Divergent Evolutionary Histories of Two Sympatric Spruce Bark Beetle Species

Molecular Ecology (2013) 22, 3318–3332 doi: 10.1111/mec.12296

Divergent evolutionary histories of two sympatric spruce bark beetle species

CORALIE BERTHEAU,* HANNES SCHULER,* WOLFGANG ARTHOFER,† DIMITRIOS N. AVTZIS,‡ 1 1 FRANCßOIS MAYER,§ SUSANNE KRUMBOCK,*€ YOSHAN MOODLEY¶ and CHRISTIAN STAUFFER* *Department of Forest and Soil Sciences, Institute of Forest Entomology, Forest Pathology and Forest Protection, Boku, University of Natural Resources and Life Sciences, Vienna, Austria, †Molecular Ecology Group, University of Innsbruck, Innsbruck, Austria, ‡Forest Research Institute, NAGREF, Vasilika, Thessaloniki, Greece, §Lutte Biologique et Ecologie Spatiale, Universite Libre de Bruxelles, Brussels, Belgium, ¶Department of Integrative Biology and Evolution, Konrad Lorenz Institute of Ethology, University of Veterinary Medicine, Vienna, Austria

Abstract Ips typographus and Pityogenes chalcographus are two sympatric Palearctic bark beetle species with wide distribution ranges. As both species are comparable in biology, life history, and habitat, including sharing the same host, Picea abies, they provide excellent models for applying a comparative approach in which to identify common historical patterns of population differentiation and the influence of species-specific ecological characteristics. We analysed patterns of genetic diversity, genetic structure and demographic history of ten I. typographus and P. chalcographus populations co-distributed across Europe using both COI and ITS2 markers. Rather than similarities, our results revealed striking differences. Ips typographus was characterised by low genetic diversity, shallow population structure and strong evidence that all extant haplogroups arose via a single Holocene population expansion event. In contrast, genetic variation and structuring were high in P. chalcographus indicating a longer and more complex evolutionary history. This was estimated to be five times older than I. typographus, beginning during the last Pleistocene glacial maximum over 100 000 years ago. Although the expansions of P. chalcographus haplogroups also date to the Holocene or just prior to its onset, we show that these occurred from at least three geographically separated glacial refugia. Overall, these results suggest that the much longer evolutionary history of P. chalcographus greatly influenced the levels of phylogeographic subdivision among lineages and may have led to the evolution of different life-history traits which in turn have affected genetic structure and resulted in an advantage over the more aggressive I. typographus. Keywords: comparative phylogeography, COI, Ips typographus, ITS2, Pityogenes chalcographus, species-specific characters

Received 25 June 2012; revision received 7 February 2013; accepted 13 February 2013

Comparative phylogeography is a powerful approach Introduction to match historical patterns of gene ﬂow, divergence An important issue in evolutionary biology is whether and speciation mechanisms among co-distributed taxa general and predictable relationships exist between the that overlap in space and time, but which are indepen- phylogeographic structure of species, their environmental dently confronted with the same historical events, requirements and species-speciﬁc ecology (Avise 2000). submitted to the same or different ecological processes and presenting similar or distinct intrinsic life-history Correspondence: Coralie Bertheau, Fax: +43 1 3686352/97; traits (Taberlet et al. 1998; Avise 2000; Hickerson et al. E-mail: [email protected] 2010). This multispecies approach offers a deeper 1Equally contributing senior authors. understanding of the evolutionary processes affecting

© 2013 John Wiley & Sons Ltd COMPARATIVE GENETIC STUDY OF BARK BEETLES 3319 phylogeographic patterns among sympatric species and develop in other Pinaceae species (Fuhrer€ & (Bermingham & Moritz 1998; Zink 2002). Concordant Muehlenbrock 1983; Bertheau et al. 2009a,b). Conse- geographic distributions among species lineages have quently, after strong storms across Europe in 1990, enabled the detection of climatic refugia, postglacial P. chalcographus was responsible for the destruction of re-colonization routes or zones of contact (Taberlet et al. eight million m3 timber of spruce and other conifers, 1998; Hewitt 2000), while unrelated phylogeographic whereas the monophagous I. typographus destroyed patterns among species highlighted the influence of four times that amount exclusively in spruce (Gregoire environmental factors, life-history and/or ecological & Evans 2004). traits which affect dispersal capacities, host specializa- The phylogeography of both species has been rela- tion or adaptability to new environments (Avise et al. tively well investigated in Europe (see Avtzis et al. 2012 1987; Bowen & Avise 1990; Peterson & Dennot 1998). for review). Ips typographus has been the subject of sev- Although the number of comparative genetic studies on eral studies using a variety of molecular markers. An insects has increased in recent years, most deal with original analysis of a fragment of the mitochondrial host-parasite interactions (Whiteman et al. 2007; Ren gene cytochrome c oxidase subunit I (COI) from 18 et al. 2008; Borer et al. 2012; among others) or closely spruce populations revealed only eight haplotypes but related species (Solomon et al. 2008; Papadopoulou et al. identified two potential glacial refugia—one in the 2009; Morgan et al. 2011; among others), and few have south Alps and the other in the Moscow region (Stauffer been carried out in Europe (Brouat et al. 2004; Hayward et al. 1999). Contrary to knowledge that spruce recolon- & Stone 2006; Kerdelhue et al. 2006; Espındola & ized Scandinavia from a refuge on the Russian plain Alvarez 2011). (Tollefsrud et al. 2008), Stauffer et al. (1999) concluded Here, we present a comparative study of two sympatric that I. typographus re-colonized Scandinavia only from insect species with similar life histories, biology and southern Europe because Russian and Lithuanian popu- habitats to determine how these have been influenced lations did not share haplotypes with Scandinavia. by their evolutionary histories. Ips typographus (L.) and Nuclear allozyme and microsatellite studies highlighted Pityogenes chalcographus (L.) are two Palearctic scolytid low diversity and lack of structure in I. typographus beetles belonging to the tribe Ipini. They have a wide populations due to high gene flow (Stauffer et al. 1999; distribution range concordant with the distribution of Salle et al. 2007; Gugerli et al. 2008). However, the their main host, the Norway spruce, Picea abies (L. H. recent discovery of cryptic nuclear copies (numts) that Karst.) (Pfeffer 1995). The two species specialize in appear very similar to authentic mitochondrial DNA exploiting weakened spruce trees especially after (mtDNA) in I. typographus (Bertheau et al. 2011) created extreme climatic events, such as storms, snow breakage doubts in the interpretation of Stauffer et al. (1999). Avt- and droughts. These events favour population growth zis et al. (2008) studied P. chalcographus in 39 European that can lead to extensive ecological and economical spruce populations, sequencing almost the complete damage, making the beetles the most serious pests to COI gene, and discovered 58 haplotypes that were par- spruce forests in Europe (Gregoire & Evans 2004). titioned into three major clades. Two of these clades are Among the reasons for their breeding success are their thought to have diverged 70 000–100 000 years ago efficient pheromone-mediated infestation and aggrega- (kya) from central to northeastern European refugia and tion behaviour (Byers 2004), their strong association are now found in northern and central Europe with with blue stain fungus (Kirisits 2004; Six & Wingfield unidirectional reproductive incompatibility (Fuhrer€ 2011) and their flexible reproductive cycles which allow 1976; Avtzis et al. 2008). Southern Europe comprised for up to three generations in particularly warm years the remaining genetic variation with glacial refugia sug- (Jurc et al. 2006; Jonsson€ et al. 2011). Both species are gested in the Apennines and the Dinaric Alps. polygamous with an endophytic life cycle as they bore While much work has been carried out, the numt galleries into the bark of their host where larval devel- problem, the lack of a fully comparative assessment opment and most often adult maturation take place. and of an informative nuclear DNA (nuDNA) marker They are mostly found together on the same trees currently preclude a common synthesis of bark beetle (Bertheau et al. 2009a). Ips typographus is confined to the phylogeography in Europe. mid- and lower parts of the trunk (Hedgren 2004; Ips typographus and P. chalcographus share many eco- Wermelinger 2004), whereas P. chalcographus segregates logical and life characteristics, but they show variable preferentially into the thinner bark of the tree’s upper levels of specialization to P. abies. As the geographic reaches and branches but it is not rare to find it in the distribution of phytophagous insects is necessarily thicker bark (Grunwald€ 1986). While both species embedded within the range of their host plants preferentially infest P. abies, only P. chalcographus is (Simonato et al. 2007), one might expect the genetic considered oligophagous, with the ability to colonize structure of the monophagous I. typographus to more

© 2013 John Wiley & Sons Ltd 3320 C. BERTHEAU ET AL. closely resemble that of P. abies than the oligophagous Materials and methods P. chalcographus. Dres & Mallet (2002) showed that host specificity and host availability play a major role in the Sampling partitioning of genetic variation and structure in phytophagous insects. Given a fully comparative approach Adults of I. typographus and P. chalcographus were col- and considering previous findings, one would expect lected at ten different locations from ten countries in con- that populations of a monophagous species would be tinental Europe covering the natural distribution area of more differentiated than an oligophagous species their main host P. abies (Table 1, Fig. 1). Between 1995 because the scarce and patchy distribution of its specific and 2010, 394 I. typographus and 476 P. chalcographus host may lead to population isolation and a reduction specimens were sampled from P. abies trees. Only one of gene flow (Peterson & Dennot 1998; Kelley et al. individual per mother gallery-the gallery system con- 2000). Alternatively, the rarity of appropriate host structed by a single female after mating-was taken to plants may force specialized I. typographus individuals prevent the sampling of siblings. All beetles were stored to disperse in search of new suitable hosts, leading to in absolute ethanol at 20 °C. À higher gene flow with a reduction of species-wide genetic structure (Lieutier 2002). To test these hypothe- DNA protocols ses, we employed a multispecies-multi-marker strategy for the combined analysis of phylogeographic patterns DNA was extracted from whole beetles for a subset of in I. typographus and P. chalcographus. We sampled both 15-48 individuals from each population using the GenE- species extensively at ten key locations and quantified lute Mammalian Genomic DNA miniprep kit (Sigma) genetic diversity, structure and demographic history according to the manufacturer’s protocol. Mitochondrial using the mitochondrial COI and nuclear internal COI gene was amplified via PCR using the sense primer transcribed spacer two region (ITS2) fragments. This described by Juan et al. (1995) and the anti-sense primer approach provided the unique opportunity of unravel- UEA10 (Lunt et al. 1996) for I. typographus (629 bp) and ling two distinctive evolutionary trajectories. We use the sense primer PcCOIF 5′- ATTATTAACAGACCG this information in combination with that of host spe- AAACG-3′ and the anti-sense primer UEA10 (Lunt et al. cies, P. abies, to discuss the relative benefits conferred 1996) for P. chalcographus (950 bp). The full ITS2, includ- through differences in life-history strategies over evolu- ing the end of the 5.8S and the beginning of the 28S tionary time and how these may have affected genetic ribosomal gene, was amplified using standard oligonu- structure. cleotide primers ITS2F/ITS2R (Campbell et al. 1993) for

Table 1 Sampling sites, abbreviations of Ips typographus and Pityogenes chalcographus populations, year of capture and geographical coordinates

Species Country Location Code Date Latitude Longitude

Ips typographus Austria Rothwald It-AtRo 1996 47° 45′N 15° 04′E Croatia Vrhovine It-CrVr 2009 44° 51′N 15° 25′E Finland Joensuu It-FiJo 2009 62° 35′N 29° 45′E France Dole^ It-FrDo 2007 47° 05′N 05° 29′E Greece Drama It-GrDa 2010 41° 08′N 24° 09′E Italy Abetone It-ItAb 2009 44° 08′N 10° 39′E Poland Hajnowka It-PlHa 2010 52° 44′N 23° 34′E Romania Belisß It-RoBe 1996 46° 04′N 23° 01′E Russia Moscow It-RuMo 1995 55° 45′N 37° 36′E Sweden Hogberget€ It-SwHo 1996 60° 27′N 15° 05′E Pityogenes chalcographus Austria Rothwald Pc-AtRo 2004 47° 45′N 15° 04′E Croatia Saborsko Pc-CrSa 2009 44° 59′N 15° 28′E Finland Jarvenpa€a€ Pc-FiJa 2004 60° 28′N 25° 06′E France Dole^ Pc-FrDo 2007 47° 05′N 05° 29′E Greece Drama Pc-GrDa 2004 41° 08′N 24° 09′E Italy Abetone Pc-ItAb 2009 44° 08′N 10° 39′E Lithuania Vilnius Pc-LiVi 2004 54° 04′N 25° 20′E Romania Bistra Pc-RoBi 2004 46° 22′N 23° 06′E Russia Sverdlovsk Pc-RuSv 2009 58° 53′N 61° 51′E Sweden Overkalix Pc-SwOv 2004 66° 19′N 22° 50′E

(A)

(C)

(B)

(D)

Fig. 1 Geographical distribution of mitochondrial haplogroups (COI) and nuclear alleles (ITS2) of two European spruce bark beetles. (A) Geographical mapping of the three COI haplogroups and the four ITS2 alleles detected in Ips typographus populations and (B) of the six haplogroups and ten ITS2 alleles detected in Pityogenes chalcographus populations. Population frequencies are approximated by the area of the circle. Haplotype and allele distributions are accompanied by Bayesian skyline plots showing changes in effective population size over evolutionary time. MtDNA COI sequences of the I. typographus (C) and P. chalcographus (D) were subjected to Bayesian skyline analysis using the software BEAST v. 1.6.1. The dashed lines correspond to the 95% conﬁdence limits while the full lines correspond to the median value. Populations are abbreviated according to Table 1. both species. Reactions were carried out in 25 lL total and 0.2 lM of the COI and ITS2 primers, respectively, 1 volumes, containing 19 the reaction buffer provided unit of Taq polymerase (Fermantas) and ~50 ng template with the polymerase, 2 mM MgCl2, 100 lM dNTPs, 0.4 DNA. PCR was performed with an initial denaturation

© 2013 John Wiley & Sons Ltd 3322 C. BERTHEAU ET AL. step at 94 °C for 3 min followed by 30 cycles of using the model of nucleotide substitution that best fits 94 °C-60 s, 48 °C (COI) or 51 °C (ITS2)-60 s, 72 °C-90 s, the data under the hierarchical likelihood-ratio test followed by a final extension at 72 °C for 10 min. PCR (hLRT) criterion, determined with Modeltest v3.7 products were purified using the GenElute PCR Clean- (Posada & Crandall 1998). The most appropriate models Up kit (Sigma-Aldrich) and sequencing was performed of COI sequence evolution were HKY+G (Hasegawa et al. at the Cancer Research Center DNA Sequencing Facility 1985) for I. typographus and K81+G+I (Kimura 1981) for (University of Chicago, Chicago, IL, USA). COI and ITS2 P. chalcographus. For the ITS2 locus, we used the JC69 sequences were visualized using Bioedit v7.0.5. (Hall model (Jukes & Cantor 1969) for both species. We used a 1999) and aligned using Clustal W (Thompson et al. Yule process tree prior, with base frequencies estimated 1994) as implemented in Bioedit. and the gamma distribution described by ten categories. Due to the recent detection of cryptic numts in We performed 350 million MCMC simulations, logging I. typographus, only haplotypes with clear, unambiguous to file every 350 000 iterations and discarded 10% as sequence chromatograms and confirmed by sequencing- burn-in. A likelihood-ratio test (LRT, Felsenstein 1988) independent PCR amplicons were used for phylogenetic supported a molecular clock model for both I. typogra- analysis (see Bertheau et al. 2011). For P. chalcographus, phus (v2 = 8.77, d.f. = 28, P < 0.05) and P. chalcographus all sequences were free from ambiguous sites and pre- (v2 = 71.15, d.f. = 119, P < 0.05). COI and ITS2 sequences mature stop codons, consistent with true mitochondrial of Ips cembrae (GenBank KC514452, KC514464), Ips origin (see Zhang & Hewitt 1996 for review). amitinus (KC514451, KC514463), Pityogenes bidentatus The ITS2 chromatograms of most individuals of (KC514453, KC514465) and Pityogenes quadridens I. typographus (89.6%) displayed numerous double peaks (KC514454, KC514466) were used as outgroups. We dated at four ambiguous positions and continuously from the the clock-like COI phylogenies by applying a global 417th nucleotide until the end of the fragment due to mutation rate of 1.87%/Myr, which is an average of three insertion/deletion polymorphisms; 58% of the individu- recently calibrated Coleopteran COI mutation rates rang- als of P. chalcographus showed double peaks at ten ing from 2.5 to 1.5%/Myr (Borer et al. 2010). Mutation ambiguous positions. These double peaks suggested the rates for the ITS2 locus do not currently exist for the Cole- superposition of two sequences reflecting heterozygos- optera, precluding the dating of nuclear phylogenies. ity. The different allele sequences were reconstructed by Statistical parsimony networks were computed, for comparing chromatograms for the forward and reverse both mitochondrial and nuclear markers, using TCS primers following Flot et al. (2006). A confirmation of version 1.21 (Clement et al. 2000). To solve the few the alleles was obtained by cloning PCR products of cladogram ambiguities that occurred, we used topologi- four heterozygote I. typographus individuals from three cal, geographical and frequency criteria (Crandall & populations (Austria, France and Italy) and ten hetero- Templeton 1993). zygote P. chalcographus individuals from six populations We investigated geographical structure in our data (Croatia, Finland, Greece, Italy, Lithuania and Roma- sets by conducting spatial analyses of molecular vari- nia). PCR products were ligated into the pTZ57R/T vec- ance (SAMOVA 1.0, Dupanloup et al. 2002), which identify tor (Fermentas) and transformed into competent JM109 groups of populations that are geographically homoge- Escherichia coli cells according to the instructions of the neous and maximally differentiated. The method manufacturer. Plasmid DNA of two to eight clones per requires the a priori definition of the number of groups individual was sequenced using the universal M13 for- (K) of populations that exist and generates F-statistics ward primer. (FSC, FST and FCT) using an AMOVA approach. The genetic differentiation among groups, expressed by the F value associated with the K groups, was computed Data analyses CT using pairwise differences between DNA sequences as

Gene diversity Hd, nucleotide diversity p and mean molecular distance (Dupanloup et al. 2002). The number of pairwise differences were calculated using program was run for 10 000 permutations from 100 Arlequin 3.5 (Excoffier & Lischer 2010) for both COI random initial conditions for two to nine differentiated and ITS2 sequences. Allelic richness r was computed for groups (K = 2 to K = 9). Occurrence of significant phy- both species populations and for both COI and ITS2 logeographic structure was also assessed by testing markers using the rarefaction method proposed by Petit whether Gst (coefficient of genetic variation over all et al. (1998) using Contrib (http://www.pierroton.inra.fr/ populations) was significantly smaller than Nst (equiva- genetics/labo/Software/Contrib/). lent coefficient taking into account the similarities Bayesian phylogenies were reconstructed using the between haplotypes) fusing 1000 permutations (see Pons software BEAST 1.6.1 (Drummond & Rambaut 2007) for & Petit 1996) in the program Permut (http://www.pierr both mtDNA and nuDNA data. Trees were generated oton.inra.fr/genetics/labo/Software/Permut/).

Population historical events were inferred using both very similar patterns of relatedness. MtDNA trees com- mitochondrial and nuclear data sets. The frequency- prised two major lineage splits at the point of species based indicators of a population expansion (or selection coalescence, yet slower evolving nuDNA was able to in non-neutral markers) Tajima’s D (Tajima 1989) and distinguish ancestral alleles B and I in I. typographus Fu’s Fs (Fu 1997) were calculated with Arlequin 3.5 and P. chalcographus respectively, and a more derived (Excoffier & Lischer 2010). Finally, we used coalescent clade (Fig. 2). Haplotype networks resolved three major simulations in BEAST 1.6.1 to reconstruct extended Bayes- mtDNA haplogroups in both species, each of which ian skyline plots, which uses the coalescent to model consisted of one to three haplotypes at high frequency, changes in the effective population size. As with the with all other haplotypes radiating mainly from these dating of phylogenetic lineage splits, we were only able by one or two base pairs (Fig. 3). In P. chalcographus to perform these analyses for the COI locus, due a gen- one of the three mtDNA haplogroups (Pc-III) was struc- eral lack of ITS mutation rates. We performed these tured into four subgroups, two of which occur at low analyses for each species using 350 million MCMC frequency (Fig. 2B, 3B). Nuclear allele networks showed iterations, logging to file every 350 000 iterations and that both species consisted of ancestral and derived discarding 10% as burn-in. clades (Fig. 3C,D). We dated the time to the most recent common ancestor (tMRCA) of all mitochondrial haplogroups to Results ~19 kyr (CI95%, 11-28 kyr) in I. typographus and to 101 kyr (CI95%, 68-132 kyr) in P. chalcographus (Table 3). Genetic diversity Divergence times were consistent whether both or only A homologous 564 bp fragment of the COI mtDNA a single outgroup sequence was used. The tMRCAs for gene was analysed in both species. Among the 394 each haplogroup varied considerably within each spe- I. typographus individuals sequenced, 22 transitions and cies, however, only those among I. typographus could three transversions resulted in 30 haplotypes (GenBank conceivably (with 95% confidence) have evolved during ITU82589, AF036150, AF036156, JN133853-JN133879; the Holocene (in the last 12 kyr). Table S1, Supporting information). The two most common were shared by 222 and 110 individuals, respec- Species geographic structure tively (Table S1, Supporting information). Total gene diversity was 0.60 0.02, while nucleotide diversity was The distribution of genetic diversity was also strikingly Æ 0.0013 0.0010. Contrastingly, mitochondrial genetic similar in both species. Generally, western Europe Æ diversity among the 476 P. chalcographus individuals was tended to be more diverse than either northern or south- higher revealing 61 transitions, 18 transversions and 94 eastern Europe in both species (Table 2). Northern and haplotypes (KC514357-KC514450, Table S2, Supporting southeastern European populations consisted of a single, information). While haplotype sharing was also high in but different mtDNA haplotype for I. typographus or this species (two haplotypes were shared by 119 and haplogroup for P. chalcographus at high frequency. 132 individuals, Table S2, Supporting information), In I. typographus, mtDNA haplogroup It-A was more gene diversity (0.87 0.05) and nucleotide diversity widely represented in north and central Europe, It-B Æ (0.012 0.006) were significantly higher in P. chalcogra- was distributed at highest frequency in Croatia and Æ phus than in I. typographus. Nuclear variation within the Greece, whereas It-C appeared at lower frequencies ITS2 region showed similar intraspecific trends, but with across Europe (Fig. 1A, Table S1, Supporting informa- decreased diversity. Four alleles were identified among 96 tion). Accounting for geography, a SAMOVA returned a

I. typographus individuals (KC514467-KC514470, Table S1, maximally significant FCT statistic (0.497, P < 0.05) for Supporting information) with a gene diversity of the two-lineage phylogenetic hypothesis which 0.51 0.01 and a nucleotide diversity of 0.003 0.002, decreased consistently for all K > 2 (Fig. S1A, Support- Æ Æ whereas 95 P. chalcographus individuals returned ten ing information). One of the two groups was made up alleles (JQ066307, JQ066310, KC514455–KC514462; Table of Croatian and Greek populations, and the second S2, Supporting information) with a gene diversity of comprising all other populations. However, Gst (0.39) 0.57 0.02 and a nucleotide diversity of 0.001 0.001. and Nst (0.35) did not differ significantly, possibly Æ Æ because of the close relationships among the three major haplogroups. Even lower variation at the ITS2 Relatedness and divergence was unable to resolve geographic structure, with the Bayesian phylogenetic trees (Fig. 2) and haplotype/ Gst ( 0.050) not differing significantly from Nst À allele networks (Fig. 3) were reconstructed for mito- ( 0.052). This is likely because alleles A and B were À chondrial and nuclear markers. Both species showed both widely distributed at high frequency across most

(A) (B)

(C)

(D)

Fig. 2 Mitochondrial and nuclear Bayesian phylogenetic trees of the two European spruce bark beetles. (A) Ips typographus COI, (B) Pityogenes chalcographus COI, (C) I. typographus ITS2, (D) P. chalcographus ITS2. All branches shown on the trees are present in 100% of the posterior sample. Clock-like mtDNA phylogenies (A and B) were linearized and dated by applying an averaged coleopteran COI mutation rate of 1.87%/Myr (calculated from Borer et al. 2010).

of Europe, and alleles C and D occurred at low fre- the three most common and star-shaped haplogroups, quency only in France and Sweden, respectively. Pc-I and Pc-IIIa were distributed in opposing north-south The higher genetic diversity of P. chalcographus gave gradients, being most frequent in northern/eastern rise to a more complex phylogeographic structure, Europe, and in Greece/western Europe respectively. although still very similar to that of I. typographus. Of The third major haplogroup Pc-IIId was at highest

(A) (B)

(C)

(D)

Fig. 3 Mitochondrial and nuclear networks of the two European spruce bark beetles. (A) Ips typographus COI, (B) Pityogenes chalcographus COI, (C) I. typographus ITS2, (D) P. chalcographus ITS2. Each circle corresponds to one haplotype; circle size gives the proportion of individuals belonging to the haplotype. Each link between circles indicates one mutational event. Populations are abbreviated according to Table 1. frequency in the Dinaric Alps of Croatia, but also mitochondrial data set was therefore signiﬁcantly geo- occurred at low frequency in western Europe (Fig. 1B). graphically structured [Gst (0.12) < Nst (0.17), P < 0.01]

The remaining less frequent haplogroups (II, IIIb and and a SAMOVA returned a maximum FCT value (0.339, IIIc) were all endemic to the Apennine population of P < 0.05) for K = 3 (Fig. S1A, Supporting information). Italy (Figs 1B and 3B), despite each being of indepen- This equated to one group composed of the Croatian dent phylogenetic origin (Fig. 2B). The P. chalcographus population, the second group with the diverse Italian

© 2013 John Wiley & Sons Ltd 3326 C. BERTHEAU ET AL. population and the third group comprising all other resulted from a recent divergence, and an early- to mid- populations. Similar to I. typographus, the less diverse Holocene population expansion, whereas P. chalcogra- nuDNA sequences were more homogeneously distrib- phus was structured into six haplogroups, was highly uted across Europe with the Gst (0.020) and Nst diverse and underwent a longer, more complex evolu- ( 0.001) not significantly different. A SAMOVA, however, tionary history. À returned a maximum and significant FCT value (0.089, P < 0.05) for K = 3 (Fig. S1B, Supporting information), Diversity partitioned similarly to the mitochondrial groups above, the exception being that one group consisted of Greece Lower genetic diversity in I. typographus supports previ- rather than Croatia. This result supports the observation ous findings (Stauffer et al. 1999; Avtzis et al. 2008), that the two most common ITS2 alleles occur at highest although the much larger sample sizes used in the pres- frequency in Italy (allele I) and Romania and Greece ent study considerably increase the number of detected (allele II). haplotypes-I. typographus 30 vs. 8; P. chalcographus 94 vs. 56-allowing for higher resolution evolutionary inference. Conversely, the greater diversity observed in the pres- Population history ent study may be attributable to the presence of numts Overall the significant values for both Tajima’s D and (Bensasson et al. 2001; Song et al. 2008). We argue Fu’s Fs statistics (Table 2) point to a hypothesis of his- against this possibility because unlike in I. typographus torical population expansions in both species. However, where numts could potentially be widespread, while overall Tajima’s D was significant (P < 0.05) for nuclear copies of mtDNA have never been identified in P. chalcographus, none of the sampled populations P. chalcographus (Arthofer et al. 2010), yet mtDNA of showed highly negative or highly significant values this species is about ten times more diverse than in (Table 2). This, in contrast to several I. typographus pop- I. typographus (Table 2). Furthermore, even if numts ulations with highly significant population expansion have contaminated the present study, the three main signatures (Table 2), suggests that demographic events haplotypes observed here (HTI, HTII and It1) were pre- shaping population structure in P. chalcographus have viously validated as originating strictly from mitochon- occurred earlier than in I. typographus. Negative values dria (Bertheau et al. 2011). for both D and Fs for the less diverse nuclear ITS2 sequences also lend support to a population expansion Evolutionary history hypothesis in both species, although only Fs was significant in I. typographus. We used extended Bayesian sky- The genetic structure of these bark beetle species line simulations to reconstruct the changes in effective reflects both their origins and their demographic histo- population size occurring through evolutionary time ries. Low variation at mtDNA and nuDNA in I. typogra- and found that while both species were characterised by phus suggests high gene flow in this species and is a rapid increase in effective population size, the I. typog- consistent with previous allozyme and microsatellite raphus expansion occurred more recently, dating back to results (Stauffer et al. 1999; Salle et al. 2007). Despite this the Holocene (~7-15 kya, Fig. 1C). In contrast, popula- low diversity, we found that I. typographus populations tion expansions among P. chalcographus haplogroups are indeed structured across Europe. However, the occurred prior to the onset of the Holocene, between 18 three star-shaped mtDNA haplogroups were linked to and 26 kya (Fig. 1D). Both population expansion events each other by single mutations and one to four step postdate haplogroup divergence times. While Bayesian mutations were found to link all 30 I. typographus haplo- skyline plots may be sensitive to heavily structured pop- types. The two most common nuclear ITS2 alleles ulations (Ho & Shapiro 2011), employment of this (A and B) were both distributed at high frequency in method is reasonable in the case presented here, given almost all populations. The two main mtDNA haplo- the relatively low level of structure observed in both groups (It-A and It-B) both evolved at or just before the study species and recent species expansion events. end of the last glacial maximum, and expanded in the Holocene (Fig 1C). We suggest the most parsimonious explanation for these results is a very recent species ori- Discussion gin and expansion from a single late-Pleistocene glacial Our comparative study of two sympatric bark beetle spe- refuge as it is unlikely that haplogroups would remain cies across European forests revealed striking differences so closely related if they have been isolated in two or in genetic variation, population structure, evolutionary more refugia. The three closely related haplogroups and demographic history. In general, I. typographus must therefore have evolved due to postexpansion exhibited low genetic diversity, shallow structure, which demographic processes. The differing frequencies of

Table 2 Indices of genetic diversity per species and molecular marker, Tajima’s D and Fu’s Fs statistics. Populations are abbreviated according to Table 1

COI ITS2

Sp. Pop. N #HT H SD r MNPD SD p SD Tajima’ D Fu’s Fs N #Al H SD r MNPD SD p SD Tajima’ D Fu’s Fs d Æ Æ Æ d Æ Æ Æ

Ips It-AtRo 42 11 0.57 0.09 4.158 [15] 0.84 0.61 0.0015 0.0012 1.73* 8.59*** 10 2 0.52 0.05 0.99 [6] 1.57 0.98 0.003 0.002 1.51 NS 4.21 NS Æ Æ Æ À À Æ Æ Æ typographus It-CrVr 47 3 0.20 0.07 1.118 [15] 0.36 0.36 0.0006 0.0007 0.98 NS 0.31 NS 10 2 0.53 0.04 0.99 [6] 1.58 0.98 0.003 0.002 1.57 NS 4.40 NS Æ Æ Æ À À Æ Æ Æ It-FiJo 45 6 0.29 0.09 2.047 [15] 0.31 0.32 0.0005 0.0006 1.82** 5.25*** 10 2 0.53 0.04 0.99 [6] 1.58 0.98 0.003 0.002 1.56 NS 4.40 NS Æ Æ Æ À À Æ Æ Æ It-FrDo 34 6 0.70 0.05 3.302 [15] 1.18 0.78 0.0021 0.0015 0.55 NS 1.00 NS 10 3 0.56 0.08 1.33 [6] 1.57 0.98 0.003 0.002 0.17 NS 2.35 NS Æ Æ Æ À À Æ Æ Æ It-GrDa 36 5 0.30 0.10 2.064 [15] 0.37 0.36 0.0007 0.0007 1.51* 3.19** 10 2 0.53 0.04 0.99 [6] 1.58 0.98 0.003 0.002 1.55 NS 4.32 NS Æ Æ Æ À À Æ Æ Æ BEETLES BARK OF STUDY GENETIC COMPARATIVE It-ItAb 35 3 0.48 0.06 1.428 [15] 0.50 0.44 0.0009 0.0009 0.06 NS 0.16 NS 10 2 0.51 0.06 0.99 [6] 1.54 0.97 0.003 0.002 1.43 NS 4.07 NS Æ Æ Æ Æ Æ Æ It-PlHa 44 9 0.53 0.09 3.901 [15] 0.73 0.55 0.0013 0.0011 1.69* 6.09*** 10 2 0.53 0.04 0.99 [6] 1.58 0.98 0.003 0.002 1.57 NS 4.40 NS Æ Æ Æ À À Æ Æ Æ It-RoBe 48 5 0.40 0.08 2.276 [15] 0.43 0.40 0.0008 0.0008 1.20 NS 2.47* 10 2 0.53 0.04 0.99 [6] 1.58 0.98 0.003 0.002 1.57 NS 4.40 NS Æ Æ Æ À À Æ Æ Æ It-RuMo 15 3 0.26 0.14 2.000 [15] 0.27 0.31 0.0005 0.0006 1.49 NS 1.55* 6 2 0.55 0.07 0.99 [6] 1.64 1.04 0.003 0.002 1.44 NS 3.56 NS Æ Æ Æ À À Æ Æ Æ It-SwHo 48 4 0.12 0.06 0.938 [15] 0.13 0.20 0.0002 0.0004 1.70* 4.22*** 10 3 0.58 0.06 1.28 [6] 1.81 1.09 0.003 0.002 0.38 NS 2.82 NS Æ Æ Æ À À Æ Æ Æ À Total 394 30 0.60 0.02 0.76 0.56 0.0013 0.0010 2.09*** 3.4E+38*** 96 4 0.51 0.01 1.53 0.92 0.003 0.002 0.07 NS 1.01 NS Æ Æ Æ À À Æ Æ Æ À À Pityogenes Pc-AtRo 47 20 0.90 0.03 19.00 [47] 5.08 2.50 0.009 0.005 0.56 NS 5.26* 10 2 0.53 0.05 1.00 [9] 0.53 0.46 0.001 0.001 1.53 NS 1.39 NS Æ Æ Æ À À Æ Æ Æ chalcographus Pc-CrSa 48 21 0.92 0.02 19.71 [47] 6.32 3.05 0.011 0.006 0.32 NS 4.30 NS 10 4 0.66 0.08 2.20 [9] 0.80 0.61 0.002 0.001 0.40 NS 0.82 NS Æ Æ Æ À À Æ Æ Æ À À Pc-FiJa 47 10 0.54 0.08 09.00 [47] 2.36 1.31 0.004 0.003 1.33 NS 1.38 NS 9 4 0.64 0.07 2.06 [9] 0.75 0.58 0.001 0.001 0.44 NS 0.83 NS Æ Æ Æ À À Æ Æ Æ À À Pc-FrDo 48 17 0.84 0.05 15.81 [47] 5.29 2.60 0.009 0.005 0.54 NS 2.40 NS 9 3 0.59 0.07 1.56 [9] 0.65 0.53 0.001 0.001 0.20 NS 0.11 NS Æ Æ Æ À À Æ Æ Æ Pc-GrDa 48 9 0.55 0.08 07.92 [47] 3.50 1.81 0.006 0.004 1.06 NS 0.88 NS 10 2 0.46 0.11 1.00 [9] 0.46 0.43 0.001 0.001 0.95 NS 0.98 NS Æ Æ Æ À Æ Æ Æ Pc-ItAb 48 17 0.92 0.02 15.86 [47] 6.93 3.31 0.012 0.007 0.23 NS 0.99 NS 10 3 0.41 0.15 1.62 [9] 0.54 0.48 0.001 0.001 0.46 NS 0.41 NS Æ Æ Æ À À Æ Æ Æ À À Pc-LiVi 47 7 0.64 0.05 06.00 [47] 2.23 1.25 0.004 0.002 1.44 NS 0.61 NS 9 3 0.60 0.09 1.69 [9] 0.69 0.56 0.001 0.001 0.21 NS 0.05 NS Æ Æ Æ À Æ Æ Æ Pc-RoBi 48 8 0.74 0.03 06.90 [47] 2.71 1.46 0.005 0.003 0.77 NS 0.60 NS 10 4 0.57 0.13 2.26 [9] 0.78 0.60 0.001 0.001 0.53 NS 0.96 NS Æ Æ Æ À Æ Æ Æ À À Pc-RuSv 48 11 0.75 0.04 09.90 [47] 3.56 1.83 0.006 0.004 1.02 NS 0.37 NS 9 4 0.63 0.08 2.06 [9] 0.72 0.57 0.001 0.001 0.53 NS 0.92 NS Æ Æ Æ À À Æ Æ Æ À À Pc-SwOv 47 13 0.70 0.07 12.00 [47] 2.95 1.57 0.005 0.003 1.12 NS 2.67 NS 9 2 0.53 0.05 1.00 [9] 0.53 0.47 0.001 0.001 1.50 NS 1.32 NS Æ Æ Æ À À Æ Æ Æ Total 476 94 0.87 0.05 6.93 3.23 0.012 0.006 1.73* 24.97*** 95 10 0.57 0.02 0.65 0.50 0.001 0.001 1.53 NS 10.38*** Æ Æ Æ À À Æ Æ Æ À À

N: Number of individuals analysed; #HT: Number of haplotypes per population; #Al: Number of alleles per population; Hd: Gene diversity and its standard deviation; r: Allelic richness after rarefaction; MNPD: Mean number pairwise differences and its standard; p:Nucleotidediversityanditsstandarddeviation;NS:Nonsignificant. *P < 0.05, **P < 0.01, ***P < 0.001. 3327 3328 C. BERTHEAU ET AL. each haplogroup among locations probably reflect vary- (either the Carpathians or Bulgaria) for haplogroup ing levels of gene flow, and/or genetic drift. However, Pc-IIIa and the Russian plain for haplogroup Pc-I. The the fact that several populations contain at least two of isolation of P. chalcographus in different refugia is also the three haplogroups at high frequency makes a deter- reflected by nuclear alleles. All refugia are presumed to mination of its origin difficult. Furthermore, shallow have contained nuclear alleles I and II, the latter of genetic structure of I. typographus is incongruent with which mutated independently in different refugia to that of its host P. abies, which is highly structured give rise to the derived nuclear clade. Thus, the genetic across Europe (Tollefsrud et al. 2008). These findings structure of the oligophagous P. chalcographus appears corroborate the view of Salle et al. (2007) who used more congruent with its host species than does that of microsatellite markers to show that local insect– the monophagous I. typographus, probably due to the tree relationships and co-adaptations are unlikely to be former species’ longer evolutionary history. According the main forces shaping the evolutionary history of to Tollefsrud et al. (2008), P. abies in southern refugia I. typographus. expanded rapidly at the advent of the Holocene, recon- Conversely, the high mtDNA genetic variation found necting western European refugia by 9 kya, and expan- within P. chalcographus is structured into six haplo- sion over the entire northern range from a single refuge groups, consistent with Avtzis et al. (2008), but our data on the Russian plain was complete by 6 kya, with both show that the two most frequent haplogroups (Pc-I and populations coming into present day secondary contact Pc-IIIa) are also distributed as far east as European Rus- in Poland. This is compatible with the expansions of sia. We also find a reduction in mtDNA diversity of Pc-IIId from Apennines, Pc-IIIa from central European eastern relative to western European populations refugia, and with Pc-I expanding out of the Russian (Fig. 1, Table 2) which suggests several different poten- plain. It is interesting to note that the two newly tial glacial refugia for this species. Comparatively, the evolved haplogroups endemic to the Apennines (Pc-IIIb coalescence of P. chalcographus mitochondrial lineages is and Pc-IIIc) do not seem to have participated in recol- five times older than I. typographus, dating back to onization. Alleles of the derived nuclear clade would ~101 kya, which is just after the peak of the last Pleisto- also be distributed across the species range through cene inter-glacial maximum, where temperatures were recolonization. even warmer than in the Holocene (Jouzel et al. 2007). The extant phylogeographic structure of P. chalcographus Species-specific life-history traits and genetic structure must therefore be the result of an initial expansion from the species origin during this last Pleistocene intergla- Pityogenes chalcographus must have shared one of its cial period, followed by glacial contraction and, finally, glacial refugia with the newly emerging I. typographus, recent expansion just prior to the Holocene (Fig. 1D). where the two species may have come under direct The most probable origin of the species is the competition with each other. However, P. chalcographus Apennines as this is the only population harbouring had already been subjected to different environments as endemic mtDNA haplotypes and it contains the ances- it established itself throughout Europe during its initial tral ITS2 allele I at highest frequency. This original Pleistocene expansion and then to severe selection pres- population, consisting of the precursors of mtDNA sure at the onset of the last Pleistocene glacial period haplogroups Pc-I/Pc-II and Pc-III as well as the ances- which funnelled its pan-European distribution into tral ITS2 allele I, then expanded at least as far east as three isolated refugia. We propose that several of the European Russia, diversifying into the mitochondrial differences in life-history characteristics between P. chal- lineages leading to haplogroups Pc-I/Pc-II, Pc-IIIa, Pc- cographus and I. typographus may have arisen, either IIIb and Pc-IIIc/Pc-IIId and simultaneously into nuclear through gain or loss of traits, in response to the harsh alleles II, IV, VIII. Of these one-step mutational variants, climate of the late Pleistocene. Whether products of only allele II is not endemic to western Europe, and its inheritance, neutral evolution or adaptation, these high frequency across the entire species suggests that it specific life traits appear to have given P. chalcographus must have evolved very soon after the initial Pleisto- an advantage over the more aggressive I. typographus. cene expansion. At ~80 kya (Table 3), a strengthening This would explain the faster postglacial dispersal of glacial period would have contracted populations to P. chalcographus (see Fig.1C,D), which would have influ- glacial refugia. Considering haplogroup/allelic distribu- enced its genetic structure and led to the contrasting tions, frequencies and the glacial distribution of P. abies phylogeographic patterns we observe between the two (Tollefsrud et al. 2008), upon which P. chalcographus is bark beetle species. One example of species-specific dependent, we propose the following glacial refugia: life-history traits that could potentially affect dispersal the Apennines for the precursors of mtDNA haplo- is the ability of P. chalcographus to over-winter at groups Pc-II, Pc-IIIb and PcIIIc/PcIIId, central Europe any ontological stage (Postner 1974), whereas only adult

Table 3 Ips typographus and Pityogenes chalcographus lineage divergence times (in kyr)

95% conﬁdence Sp. Haplogroups Distance tMRCA interval

Ips typographus Haplogroups It-A, B, C 0.00072 19 269 11° 108–28° 066 Haplogroup It-A 0.00069 18 434 10° 722–27° 164 Haplogroup It-B 0.00050 13 424 7247–21° 036 Haplogroup It-C 0.00032 8 634 3322–15° 081 Pityogenes chalcographus Haplogroups Pc-I, II, III 0.00376 100 616 68° 098–132° 249 Haplogroups Pc-I, II 0.00313 83 752 51° 559–124° 359 Haplogroups Pc-III 0.00334 89 523 58° 695–123° 535 Haplogroups Pc-IIIb, c, d 0.00314 84 051 55° 774–122° 598 Haplogroup Pc-I 0.00190 50 881 35° 166–71° 025 Haplogroup Pc-II 0.00122 32 742 12° 819–55° 503 Haplogroup Pc-IIIa 0.00204 54 737 35° 273–73° 736 Haplogroup Pc-IIIb 0.00164 44 025 22° 457–69° 100 Haplogroup Pc-IIIc 0.00117 31 229 14° 486–52° 750 Haplogroup Pc-IIId 0.00233 62 268 34° 831–94° 552 tMRCA: The most recent common ancestor.

I. typographus are able to survive a winter (Austara& lation expansion or contraction, may have been transmit- Midtgaard 1986; Christiansen & Bakke 1988; Hrasovec ted via introgression and divergent selection to other et al. 2011). This may also explain the loss of large tracts areas upon secondary contact, as was recently demon- of spruce forests due to I. typographus in recent years, as strated in Heliconius butterflies (Dasmahapatra et al. 2012; less severe European winters tend to promote popula- Nadeau et al. 2012). This would confer unto a species tion explosions (Jonsson€ et al. 2009). Additionally, the such as P. chalcographus, the capacity to shift on alterna- ability of P. chalcographus to colonize other Pinaceae tive host species, while maintaining the ability to exploit species and to thrive on different parts of the tree is a its natural host (Bertheau et al. 2012). life-history trait that not only limits the interspecific competition between the two species, but may also have influenced genetic structure. Host specialization has Conclusions been associated with extreme reduction in genetic The unique findings of this study highlight the potential diversity in scolytid beetles (Kelley et al. 2000). A simi- of a multispecies approach. By combining knowledge lar association with the lower genetic diversity of the on host and insect species with mitochondrial and monophagous I. typographus, relative to the high diver- nuclear DNA markers, we were able to unravel the evo- sity of the oligophagous P. chalcographus was observed lutionary histories of two prominent and economically in this study. However, this may be more attributable important bark beetle species. These results provide a to the younger evolutionary history of I. typographus platform from which the relationship between genetic rather than a reduction of its diversity due to host structure and intrinsic life-history factors may be specialization. While Kelley et al. (2000) argued that iso- defined, both physiologically and by genome-wide lated host patches tend to hinder gene flow and pro- screening for signals of positive selection and speciation mote genetic differentiation among specialized species, islands. A variation to this approach, using transcripto- we show that I. typographus conforms more to Lieutier’s mes, has already been applied to another closely related (2002) alternative hypothesis that patchy host distribu- species, the mountain pine beetle, Dendroctonus ponderosae tion would force highly specialized species to undergo (Keeling et al. 2012) yielding positive results. recurrent long-distance migration, thus increasing gene flow between populations and reducing species-wide neutral genetic structure. The converse that the ability Acknowledgements to colonize alternative host tree species promotes differentiation is unlikely to be the case for P. chalcographus, We thank the Austrian Science Fund (P21147-B17), European Territorial Co-operation Austria-Czech Republic 2007–2012, as its mitochondrial haplogroup structure was not European Union Seventh Framework Programme FP7 2007– found to correlate with host tree species (Bertheau et al. 2013 (KBBE 2009-3) under grant agreement 245268 ISEFOR and 2012). It is possible that beneficial life-history traits Federal Ministry of Agriculture, Forestry, Environment and evolved in one geographic area, whether through popu- Water Management, Austria for financial support; B. Heinze,

H. Mihalciuc, M. Pernek, N. I. Ukhova, H. Viiri, R. Wegensteiner Brouat C, Chevallier H, Meusnier S, Noblecourt T, Rasplus JY and P. Zolubas for providing populations; A. Drummond for (2004) Specialization and habitat: spatial and environmental his advice on the Bayesian skyline plots; Godfrey Hewitt, Sean effects on abundance and genetic diversity of forest general- Schoville and three anonymous referees for their comments on ist and specialist Carabus species. Molecular Ecology, 13, earlier drafts of the manuscript. 1815–1826. Byers JA (2004) Chemical ecology of bark beetles in a complex olfactory landscape. In: Bark and Wood Boring Insects in Living References Trees in Europe, A Synthesis(eds Lieutier F, Day KR, Battisti A, Gregoire JC & Evans HF), pp. 89–135. Kluwer Academic Arthofer W, Avtzis DN, Riegler M, Stauffer C (2010) Mitochon- Publishers, Germany. drial phylogenies in the light of pseudogenes and Wolbachia: Campbell JC, Steffen-Campbell JD, Werren JH (1993) Phylog- re-assessment of a bark beetle dataset. ZooKeys, 56, 269–280. eny of the Nasonia species complex (Hymenoptera: Pteromal- Austara Ø, Midtgaard F (1986) On the longevity of Ips typogra- idae) inferred from internal transcribed spacer (ITS2) and phus adults. Journal of Applied Entomology, 102, 106–111. 28rDNA sequences. Insect Molecular Biology, 2, 225–237. Avise JC (2000) Phylogeography: The History and Formation of Christiansen E, Bakke A (1988) The spruce bark beetle of Eur- Species. Harvard University Press, Cambridge, MA, 447 p. asia. In: Dynamic of Forest Insect Population; Patterns, Causes, Avise JC, Arnold J, Ball RM et al. (1987) Intraspecific phyloge- Implications(ed Berryman AA), pp. 480–505. Plenum Press, ography: the mitochondrial DNA bridge between population New York. genetics and systematics. Annual Review of Ecology and Clement M, Posada D, Crandall KA (2009) TCS: a computer Systematic, 18, 489–522. program to estimate gene genealogies. Molecular Ecology, 9, Avtzis DN, Arthofer W, Stauffer C (2008) Sympatric occurrence 1657–1659. of diverged mtDNA lineages of Pityogenes chalcographus Crandall KA, Templeton AR (1993) Empirical tests of some (Coleoptera, Scolytinae) in Europe. Biological Journal of the predictions from coalescent theory with applications to intra- Linnean Society, 94, 331–340. specific phylogeny reconstruction. Genetics, 134, 959–969. Avtzis DN, Bertheau C, Stauffer C (2012) What is Next in Bark Dasmahapatra KK, Walter JR, Briscoe AD et al. (2012) Butterfly Beetle Phylogeography? Insects, 3, 453–472. genome reveals promiscuous exchange of mimicry adapta- Bensasson D, Zhang DX, Hartl DL, Hewitt GM (2001) Mito- tions among species. Nature, 487, 94–98. chondrial pseudogenes: evolution’s misplaced witnesses. Dres M, Mallet J (2002) Host races in plant-feeding insects and Trends in Ecology and Evolution, 16, 314–321. their importance in sympatric speciation. Philosophical Trans- Bermingham E, Moritz C (1998) Molecular phylogenetics and actions of the Royal Society of London Biological Sciences, 357, evolution. Molecular Ecology, 7, 367–369. 471–492. Bertheau C, Salle A, Rossi JP et al. (2009a) Colonisation of Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolution- native and exotic conifers by indigenous bark beetles (Cole- ary analysis by sampling trees. BMC Evolutionary Biology, 7. optera: Scolytinae) in France. Forest Ecology and Management, Dupanloup I, Schneider S, Excoffier L (2002) A simulated 258, 1619–1628. annealing approach to define the genetic structure of popula- Bertheau C, Salle A, Roux-Morabito G, Garcia J, Certain G, tions. Molecular Ecology, 11, 2571–2581. Lieutier F (2009b) Preference-performance relationship and Espındola A, Alvarez N (2011) Comparative phylogeography influence of plant relatedness on host use by Pityogenes chal- in a specific and obligate pollination antagonism. PLoS ONE, cographus L. Agricultural and Forest Entomology, 11, 389–396. 6, e28662. Bertheau C, Schuler H, Krumbock S, Arthofer W, Stauffer C Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new (2011) Hit or miss in phylogeographic analyses: the case of series of programs to perform population genetics analyses the cryptic NUMTs. Molecular Ecology Resources, 11, 1056– under Linux and Windows. Molecular Ecology Resources, 10, 1059. 564–567. Bertheau C, Bankhead-Dronnet S, Martin C, Lieutier F, Roux- Felsenstein J (1988) Phylogenies from molecular sequences: infer- Morabito G (2012) Lack of genetic differentiation after host ences and reliability. Annual Review of Genetics, 22, 521–565. range extension argues for the generalist nature of Pityogenes Flot JF, Tillier A, Samadi S, Tillier S (2006) Phase determination chalcographus (Curculionidae: Scolytinae). Annals of Forest from direct sequencing of length-variable DNA regions. Science, 69, 313–323. Molecular Ecology Notes, 6, 627–630. Borer M, Alvarez N, Buerki S, Margraf N, Rahier M, Naisbit Fu YX (1997) Statistical tests of neutrality of mutations against RE (2010) The phylogeography of an alpine leaf beetle: population growth, hitchhiking and background selection. divergence within Oreina elongata spans several ice ages. Genetics, 147, 915–925. Molecular Phylogenetics and Evolution, 57, 703–709. Fuhrer€ E (1976) Fortpflanzungsphysiologische Unvertragli-€ Borer M, Arrigo N, Buerki S, Naisbit RE, Alvarez N (2012) chkeit beim Kupferstecher (Pityogenes chalcographus L.) – Ein Climate oscillations and species interactions: large-scale neuer Ansatz zur Borkenkaferbek€ ampfung?€ Forstarchiv, 47, congruence but regional differences in the phylogeographic 114–117. structures of an alpine plant and its monophagous insect. Fuhrer€ E, Muehlenbrock B (1983) Experimental breeding of Journal of Biogeography, 39, 1487–1498. Pityogenes chalcographus L. in different conifers. Zeitschrift fur€ Bowen BW, Avise JC (1990) Genetic structure of Atlantic and Angewandte Entomologie, 96, 228–232. Gulf of Mexico populations of sea bass, menhaden, and Gregoire JC, Evans H (2004) Damage and control of BAWBILT sturgeon: influence of zoogeographic factors and life-history organisms, an overview. In: Bark and Wood Boring Insects in patterns. Marine Biology, 107, 371–381. Living Trees in Europe, A Synthesis (eds Lieutier F, Day KR,

Battisti A, Gregoire JC & Evans HF), pp. 19–38. Kluwer Aca- Kelley ST, Farrell B, Mitton JB (2000) Effects of specialization demic Publishers, Germany. on genetic differentiation in sister species of bark beetles. Grunwald€ M (1986) Ecological segregation of bark beetles Heredity, 84, 218–227. (Coleoptera, Scolytidae) of spruce. Journal of Applied Entomol- Kerdelhue C, Magnoux E, Lieutier F, Roques A, Rousselet J (2006) ogy, 101, 176–187. Comparative population genetic study of two oligophagous Gugerli F, Gall R, Meier F, Wermelinger B (2008) Ips typogra- insects associated with the same hosts. Heredity, 97, 38–45. phus populations do not invoke genetic differentiation. Forest Kimura M (1981) Estimation of evolutionary distances between Ecology and Management, 256, 405–409. homologous nucleotide sequences. Proceedings of the National Hall TA (1999) BioEdit: a user-friendly biological sequence Academy of Sciences, 78, 454–458. alignment editor and analysis program for Windows 95/98/ Kirisits T (2004) Fungal associates of European bark beetles NT. Nucleic Acids Symposium Series, 41, 95–98. with special emphasis on the Ophiostomatoid fungi. In: Bark Hasegawa M, Kishino K, Yano T (1985) Dating the human-ape and Wood Boring Insects in Living Trees in Europe, A Synthesis splitting by a molecular clock of mitochondrial DNA. Journal (eds Lieutier F, Day KR, Battisti A, Gregoire JC & Evans of Molecular Evolution, 22, 160–174. HF), pp. 181–236. Kluwer Academic Publishers, Germany. Hayward A, Stone GN (2006) Comparative phylogeography Lieutier F (2002) Mechanisms of resistance in conifers and bark across two trophic levels: the oak gall wasp Andricus kollari beetle attack strategies. In: Mechanisms and Deployment of and its chalcid parasitoid Megastigmus stigmatizans. Molecular Resistance in Trees to Insects (eds Wagner MR, Clancy KM, Ecology, 15, 479–489. Lieutier F & Paine TD), pp. 31–77. Kluwer Academic Hedgren PO (2004) The bark beetle Pityogenes chalcographus (L.) Publishers, Dordrecht. (Scolytidae) in living trees: reproductive success, tree mortality Lunt D, Zhang DX, Szymura JM, Hewitt GM (1996) The insect and interaction with Ips typographus. Journal of Applied Ento- COI gene: evolutionary patterns and conserved primers for mology, 128, 161–166. phylogenetic studies. Insect Molecular Biology, 5, 153–165. Hewitt GM (2000) The genetic legacy of the Quaternary ice Morgan K, O’Loughlin SM, Chen B et al. (2011) Comparative ages. Nature, 405, 907–913. phylogeography reveals a shared impact of Pleistocene envi- Hickerson MJ, Carstens BC, Cavender-Bares J et al. (2010) Phy- ronmental change in shaping genetic diversity within nine logeography’s past, present, and future: 10 years after Avise Anopheles mosquito species across the Indo-Burma biodiver- 2000. Molecular Phylogenetic and Evolution, 54, 291–301. sity hotspot. Molecular Ecology, 20, 4533–4549. Ho SYW, Shapiro B (2011) Skyline-plot methods for estimating Nadeau NJ, Whibley A, Jones RT et al. (2012) Genomic demographic history from nucleotide sequences. Molecular islands of divergence in hybridizing Heliconius butterflies Ecology Resources, 11, 423–434. identified by large-scale targeted sequencing. Philosophical Hrasovec B, Kasumovic L, Franjevic M (2011) Overwintering of Transactions of the Royal Society of London Biological Sciences, eight toothed spruce bark beetle (Ips typographus) in spruce 367, 343–353. forests of north Velebit. Croatian Journal of Forest Engineering, Papadopoulou A, Anastasiou I, Keskin B, Vogler AP (2009) 32, 211–222. Comparative phylogeography of tenebrionid beetles in the Jonsson€ AM, Appelberg G, Harding S, Barring€ L (2009) Spatio- Aegean archipelago: the effect of dispersal ability and habitat temporal impact of climate change on the activity and preference. Molecular Ecology, 18, 2503–2517. voltinism of the spruce bark beetle, Ips typographus. Global Peterson MA, Dennot RF (1998) The influence of dispersal and Change Biology, 15, 486–499. diet breadth on patterns of genetic isolation by distance in Jonsson€ AM, Harding S, Krokene P et al. (2011) Modelling the phytophagous insects. The American Naturalist, 152, 428–446. potential impact of global warming on Ips typographus Petit RJ, El MA, Pons O (1998) Identifying populations for con- voltinism and reproductive diapause. Climatic Change, 109, servation on the basis of genetic markers. Conserversation 695–718. Biology, 12, 844–855. Jouzel J, Masson-Delmotte V, Cattani O et al. (2007) Orbital Pfeffer A (1995) Zentral und Westpal€aarktische Borken und and millennial Antarctic climate variability over the past Kernk€afer. Pro Entomologia, c/o Naturhistorisches Museum 800,000 years. Science, 317, 793–796. Basel, Basel, Switzerland. Juan C, Oromi P, Hewitt GM (1995) Mitochondrial DNA Pons O, Petit RJ (1996) Measuring and testing genetic differen- phylogeny and sequential colonization of Canary Islands by tiation with ordered versus unordered alleles. Genetics, 144, darkling beetles of the genus Pimelia (Tenebrionidae). Pro- 1237–1245. ceedings of the Royal Society of London B, 261, 173–180. Posada D, Crandall KA (1998) MODELTEST: testing the model Jukes TH, Cantor CR (1969) Evolution of protein molecules. In: of DNA substitution. Bioinformatics, 14, 817–818. Mammalian Protein Metabolism(ed Munro HM), pp. 21–132. Postner M (1974) Scolytidae (= Ipidae), Borkenkafer.€ In: Die Academic Press, New York. Forstsch€adlinge Europas. Band 2 (ed Schwenke W), pp. 334– Jurc M, Perko M, Dzeroski S, Demsar D, Hrasovec B (2006) 482. Parey, Berlin. Spruce bark beetles (Ips typographus, Pityogenes chalcographus, Ren Z, Zhu B, Wang D, Ma E, Su D, Zhong Y (2008) Compara- Col.: Scolytidae) in the Dinaric mountain forests of Slovenia: tive population structure of Chinese sumac aphid Schlechten- monitoring and modeling. Ecological Modelling, 94, 219–226. dalia chinensis and its primary host plant Rhus chinensis. Keeling CI, Henderson H, Li M et al. (2012) Transcriptome Genetica, 132, 103–112. and full-length cDNA resources for the mountain pine bee- Salle A, Arthofer W, Lieutier F, Stauffer C, Kerdelhue C (2007) tle, Dendroctonus ponderosae Hopkins, a major insect pest of Phylogeography of a host-specific insect: genetic structure of Ips pine forests. Insect Biochemistry and Molecular Biology, 42, typographus in Europe does not reflect the past fragmentation of 525–536. its host. Biological Journal of the Linnean Society, 90,239–246.

Simonato M, Mendel Z, KerdelhueCet al. (2007) Phylogeogra- focusing mainly on the biology and evolution of endos- phy of the pine processionary moth Thaumetopoea wilkinsoniin ymbionts in Tephritids. W.A. is assistant professor at the the Near East. Molecular Ecology, 16, 2273–2283. Molecular Ecology Group at the University of Innsbruck Six DL, Wingfield MJ (2011) The role of phytopathogenicity in and interested in all aspects of arthropod evolution and bark beetle–fungus symbioses: a challenge to the classic par- adigm. Annual Review of Entomology, 56, 255–272. symbiosis. D.N.A’s research focuses on studying host Solomon SE, Bacci M, Martins J, Goncßalves Vinha G, Mueller association of forests pests and resolving phylogeo- UG (2008) Phylogeography of leafcutter ants (Atta spp.) pro- graphic patterns within pest species, particularly bark vides new insight into the origins of Amazonian diversity. beetles, in an endeavour to identify evolutionary forces PLoS ONE, 3, e2738. that shape divergence. F.M. is a PhD candidate in Forest Song H, Buhay JE, Whiting MF, Crandall KA (2008) Many spe- Entomology working on bark beetles’ ecology and cies in one: DNA barcoding overestimates the number of genetic. S.K. is a technical assistant. Y.M. is a population species when nuclear mitochondrial pseudogenes are coam- plified. Proceedings of the National Academy of Sciences USA, and evolutionary geneticist with a broad interest in 10, 13486–13491. hybrid speciation, species-wide phylogeography and Stauffer C, Lakatos E, Hewitt GM (1999) Phylogeography and conservation. C.S.’s research focuses on the phylogeogra- postglacial colonization routes of Ips typographus L. (Coleop- phy of scolytids and tephritids, and on host-parasite tera, Scolytidae). Molecular Ecology, 8, 763–773. co-evolution. Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson JF (1998) Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology, 7, 453–464. Tajima F (1989) The effect of change in population size on DNA polymorphism. Genetics, 123, 597–601. Data accessibility Thompson JD, Higgins DG, Gibson TJ (1994) Clustal W: COI: Ips typographus GenBank accession numbers: improving the sensitivity of progressive multiple sequence ITU82589, AF036150, AF036156, JN133853–JN133879; alignment through sequence weighting, positions-specific gap penalties and w eight matrix choice. Nucleic Acids Pityogenes chalcographus: KC514357–KC514450 and out- Research, 22, 4673–4680. group species KC514451–KC514454. Tollefsrud MM, Kissling R, Gugerli F et al. (2008) Genetic con- ITS2: Ips typographus GenBank accession numbers: sequences of glacial survival and postglacial colonization in KC514467–KC514470, Pityogenes chalcographus: JQ066307, Norway spruce: combined analysis of mitochondrial DNA JQ066310, KC514455–KC514462 and outgroup species – and fossil pollen. Molecular Ecology, 17, 4134 4150. KC514463–KC514466. Wermelinger B (2004) Ecology and management of the spruce GenBank accession numbers of each COI haplotypes bark beetle Ips typographus - a review of recent research. Forest Ecology and Management, 202, 67–82. and ITS2 alleles for both species uploaded as online Whiteman NK, Kimball RT, Parker PG (2007) Co-phylogeogra- supporting information Tables S1 and S2 (Supporting phy and comparative population genetics of the threatened information). Galapagos hawk and three ectoparasite species: ecology shapes population histories within parasite communities. Molecular Ecology, 16, 4759–4773. Supporting information Zhang DX, Hewitt GM (1996) Nuclear integrations: challenges Additional supporting information may be found in the online for mitochondrial DNA markers. Trends in Ecology and Evolu- version of this article. tion, 11, 247–251. Zink RM (2002) Methods in comparative phylogeography, and Table S1 Mitochondrial haplotypes and nuclear alleles found their application to studying evolution in the North American in Ips typographus populations. Codes for populations are given aridlands. Integrative and Comparative Biology, 42, 953–959. in Table 1.

Table S2 Mitochondrial haplotypes and nuclear alleles found in Pityogenes chalcographus populations. Codes for populations This study is the result of a joint European Territorial are given in Table 1. Co-operation Austria-Czech Republic 2007–2012, EFRE Fig. S1 Plots of the FCT for different values of K, the number of project of C.S.’s and Y.M.’s institute. This work is a part population groups, for both Ips typographus and Pityogenes chal- of C.B.’s post-doctoral research focused on ecology and cographus generated using SAMOVA (Dupanloup et al. 2002). evolution of plant–insect relationships, especially conif- (A) COI marker, (B) ITS2 marker. FCT plot in solid line for erous - forest insects. H.S. is post-doctoral researcher I. typographus and in dotted line for P. chalcographus.