In this study, we document the first North American record of the gall wasp Aulacidea pilosellae Kieffer (Hymenoptera: Cynipidae), a species native to Central Europe, collected in the Ottawa Valley, Ontario, Canada. The detection of this species in Canada was possible due to (i) rapidly available public DNA barcode data and (ii) DNA sequencing of a biocontrol agent, before its release. Here we release DNA sequences and collateral information associated with our detection of A. pilosellae.

In both Canada and the United States of America, A. pilosellae is being explored as a promising candidate biocontrol agent for several species of hawkweeds in the genus Pilosella Vaillant (Asteraceae), including P. caespitosa (Dumortier) Sell and West, P. glomerata (Froelich) Arvet-Touvet, and P. piloselloides (Villars) Soják (Grosskopf et al. Reference Grosskopf, Wilson and Littlefield2008; Moffat et al. Reference Moffat, Lalonde, Ensing, De Clerck-Floate, Grosskopf-Lachat and Pither2013). As part of the host range assessment for this biocontrol programme, the lead author of this paper (C.E.M.) was investigating the potential existence of morphologically cryptic types (host races or species) of A. pilosellae, by examining sequence variation in two gene regions, cytochrome c oxidase I (COI) and 28S-D2, the D2 region of the large-subunit of RNA 28S (Moffat 2012). Through a routine comparison of these A. pilosellae sequences to all publically available DNA sequence records on GenBank (http://www.ncbi.nlm.nih.gov/genbank/), using a standard nucleotide basic local alignment search tool (BLASTn; Altschul et al. Reference Altschul, Madden, Schäffer, Zhang, Zhang and Miller1997), the COI sequence of the European A. pilosellae was found to be most similar to that of a specimen identified only to order, and differing by only one base pair (a third codon synonymous C/T substitution occurring in the fourth membrane spanning helix). Surprisingly, this Hymenoptera specimen was collected in eastern Ontario, Canada (Table 1), not in Central Europe as was expected. As no prior sequence records exist for A. pilosellae, it was expected that the most similar sequences/species determined from the BLASTn search would be one of the four other species of Aulacidea (A. freesei Nieves-Aldrey, A. hieracii Bouché, A. phlomica Belizin, or A. tragopogonis Thomson) for which GenBank records for COI sequences had already been created (DQ012627, DQ012628, DQ012629, and AY368922, respectively). Instead, the COI sequence identified only as “Hymenoptera sp.”, and collected in eastern Ontario, emerged as the most similar to the COI sequence of A. pilosellae (Table 1). While another of the European Aulacidea species, A. hieracii, has been detected in North America (Sliva and Shorthouse Reference Sliva and Shorthouse2006), we are confident that our Ontario specimen is not a member of that species as it differs substantially from A. pilosellae in both mitochondrial and nuclear markers and in body size (Moffat 2012).

Table 1 Voucher data for specimens of Aulacidea pilosellae collected in Central Europe and the Ottawa Valley, Ontario, Canada.

After verifying the identity and geographic origin of the Ontario specimen (done by C.E.M. contacting its collector, M.A.S.), this discussion led to the sequencing of one additional gene region for the Ontario specimen, the nuclear region 28S-D2. This revealed an identical sequence to that of the A. pilosellae collected in Europe (note: the 28S-D2 sequence from the Ontario specimen was identical to most specimens collected in Europe, and varied by <0.2% from all collected specimens; (Moffat 2012).

We feel that the type of communication described here exemplifies the use of DNA barcodes and a productive use of a public DNA library – in which both pre-publication and pre-taxonomic annotation data can be useful (i.e., upon discovery of the DNA sequence match, the Ontario specimen was identified only to order). While clearly there is no restriction to the use or distribution of data on GenBank (Benson et al. Reference Benson, Karsch-Mizrachi, Lipman, Ostell and Wheeler2008), this case demonstrates that the best interests of each author were served by discussion and coordination. The publication of this note should serve as an example of the productive integration of DNA barcode data by both data generator and data communities in the development, dissemination, and integration of public data sets. As exemplified here, such collaboration led to the first documentation of a European species in North America, which may affect the Pilosella hawkweed biocontrol programme by potentially decreasing the screening time before introduction. To date, we know of relatively few other published cases where a non-North American species was first documented in North America using DNA barcodes (e.g., deWaard et al. Reference deWaard, Landry, Schmidt, Derhousoff, McLean and Humble2009; Humble et al. Reference Humble, deWaard and Quinn2009; deWaard et al. Reference deWaard, Humble and Schmidt2010; Fernandez-Triana Reference Fernandez-Triana2010). We believe this is the first DNA barcode-based detection of a candidate biocontrol agent before release.

Aulacidea pilosellae is a small (1.0–1.5 mm) univoltine to bivoltine (trivoltine) cynipid, known to induce small galls, most commonly on the leaf midribs, but also on the stems and stolons, on select members of hawkweeds in the genus Pilosella (Dalla Torre and Kieffer Reference Dalla Torre and Kieffer1910; Eady and Quinlan Reference Eady and Quinlan1963; Buhr Reference Buhr1964). Aulacidea pilosellae is known from Central Europe (Ionescu Reference Ionescu1957) and the Mediterranean Basin (Houard Reference Houard1913), and is reported from Germany (Dalla Torre and Kieffer Reference Dalla Torre and Kieffer1910); France (Folliot Reference Folliot1964); Switzerland, Poland, and the Czech Republic (Moffat et al. Reference Moffat, Lalonde, Ensing, De Clerck-Floate, Grosskopf-Lachat and Pither2013); Spain (Nieves-Aldrey et al. Reference Nieves-Aldrey, Vardal and Ronquist2005); the United Kingdom (Eady and Quinlan Reference Eady and Quinlan1963); Israel (Argaman Reference Argaman1988); Romania (Ionescu Reference Ionescu1957); and Hungary (Sárospataki Reference Sárospataki1999). Across its native range, it is most commonly reported on Pilosella officinarum Vaillant (Asteraceae) (as the synonymised name Hieracium pilosella Linnaeus; Dalla Torre and Kieffer Reference Dalla Torre and Kieffer1910; Ionescu Reference Ionescu1957; Eady and Quinlan Reference Eady and Quinlan1963), but also reported from several other Pilosella species (Buhr Reference Buhr1964, Moffat et al. Reference Moffat, Lalonde, Ensing, De Clerck-Floate, Grosskopf-Lachat and Pither2013). All hawkweeds in the genus Pilosella present in North America are non-native and of European origin.

Collection of A. pilosellae in North America consisted of setting a Malaise trap (Townes Reference Townes1962) erected on a small ridge near the Bonnechere River, Ontario, Canada in April 2010 and maintained until November 2010 (Table 2). The trap was emptied every two weeks and trap contents were maintained at −20 °C. A high-resolution panoramic photograph of the collection site (taken on 2 April 2010, approximately eight weeks before the trapping event) can be viewed at http://www.gigapan.com/gigapans/46381. The unsorted sample was transferred to the University of Guelph (Ontario, Canada) where contents were sorted to order and then morphospecies. Selected specimens were photographed (Fig. 1) and then tissue sampled for DNA extraction (Table 1).

Fig. 1 The specimen of Aulacidea pilosellae (Kieffer) (ASGLE2-0266) collected in late May 2010 in the Ottawa Valley, Ontario, Canada (45.51°N, 76.977°W).

Table 2 Timeline for the collection, extraction, sequencing, and querying of the sequence from the Ottawa Valley specimen in a public database.

Sequencing of the North American A. pilosellae consisted of a DNA extract prepared from single leg using a glass-fibre extraction protocol (Ivanova et al. Reference Ivanova, deWaard and Hebert2006). The DNA extracts were re-suspended in 30 μl of dH₂O, and a the CO1 DNA barcode region (a 658-base pair region near the 5' terminus of the COI gene) was amplified using standard insect barcoding region primers LepF1 (5'-ATTCAACCAATCATAAAGATATTGG-3') and LepR1 (5'-TAAACTTCTGGATGTCCAAAAAATCA-3') following established protocols (as in Smith et al. Reference Smith, Rodriguez, Whitfield, Deans, Janzen and Hallwachs2008). The variable D2 region of the rDNA 28S region was amplified and sequenced using the primers D2B (5'-GTCGGGTTGCTTGAGAGTG-3') and D3Ar (5'-TCCGTGTTTCAAGACGGGTC-3') (Saux et al. Reference Saux, Fisher and Spicer2004). The resultant amplicons were bi-directionally sequenced. All laboratory information for the individual sequences can be retrieved from the Barcode of Life Data System (BOLD) (Ratnasingham and Hebert Reference Ratnasingham and Hebert2007), using the Process ID (sequence accession). All sequence data and detailed collection information is available on BOLD (www.barcodinglife.org) in the public data set: First Canadian Record of Aulacidea pilosellae (Cynipidae) (dx.doi.org/10.5883/DS-ASCYN1) and in Table 1.

Collections of A. pilosellae in Central Europe were made by rearing specimens from galls on whole live Pilosella plants in June 2010, as described in Moffat et al. (Reference Moffat, Lalonde, Ensing, De Clerck-Floate, Grosskopf-Lachat and Pither2013). Sequencing of Central European A. pilosellae were conducted as described in Moffat (Reference Moffat2012). Briefly, the primers LCOI490 (5'-GGTCAACAAATCATAAAGATATTGG-3') and HCO2198 (5'-TAAACTTCAGGGTGACCAAAAAATCA-3') (Folmer et al. Reference Folmer, Black, Hoeh, Lutz and Vrijenhoek1994) were used to amplify the CO1 gene region, and the primers 28S-D2(F) (5'-CGTGTTGCTTGATAGTGCAGC-3') and 28S-D2(R) (5'-TCAAGACGGGTCCTGAAAGT-3') (Heraty et al. Reference Heraty, Hawks, Kostecki and Carmichael2004; Ács et al. Reference Ács, Challis, Bihari, Blaxter, Hayward and Melika2010) were used to amplify the D2 region of the nuclear 28S rDNA gene. Sequences for both regions are available in GenBank and BOLD (Table 1).

The standard nucleotide basic local alignment search tool (BLASTn) (Altschul et al. Reference Altschul, Madden, Schäffer, Zhang, Zhang and Miller1997) was used to search nucleotide databases (all GenBank and other databases; http://blast.ncbi.nlm.nih.gov) for the most similar available sequences to the sequenced COI gene regions. We used all default search options (i.e., Database=Others (Nucleotide collection (nr/nt)); Optimize for: Highly similar sequences (megablast)).

The COI sequence from the Ontario collection, as well as the Central European collection (Table 1), is characteristic of Hymenoptera mitochondrial DNA with a high AT content (74%). BOLD automatically assigns globally unique identifiers (GUI) to specimens that are 2% divergent from other records in the database (for details on the derivation see Ratnasingham and Hebert Reference Ratnasingham and Hebert2013). These different GUI may or may not represent different species – but in either case serve as a useful “label” for specimens that do not yet have a taxonomically assigned species name. At the time of writing, searches in BOLD for this GUI (BOLD:AAU8720) yield five additional public BOLD records (JSHYM075-11, JSHYM293-11, JSHYM593-11, JSHYN760-11, JSHYP286-11) collected at a different eastern Ontario locality (north of Brockville, Ontario, Canada; 44.6214°N, 75.7734°W), indicating that our detection is not an isolated incident. These data associated with these public specimens, and others that are recovered in the future, are all retrievable from BOLD by searching the GUI BOLD:AAU8720. The CO1 DNA sequence of the Ontario specimen and associated record (Table 1) was made rapidly public following sequencing, as part of the iBOL program (ibol.org). The time delay (two months) was sufficient to perform simple validation tests regarding the ordinal identification of the sequence on the BOLD database and to rule out possible contamination, but not to assign a taxonomically valid species name (Table 2).

Part of the DNA barcoding initiative is the community-wide agreement to the policy of rapid data release pre-publication (following the Fort Lauderdale Principles – Wellcome Trust 2011). Here, in broad terms, there are “data consumers” (DC) and “data producers” (DP). Data producers agree to produce high-quality data and to make that data immediately and freely available. Data consumers agree to appropriately cite the course and acknowledge the DP’s while being free to use the data in any creative way. In particular, “… the best interests of the community are served when all act responsibly to promote the highest standards of respect for the scientific contribution of others. In some cases, this might best be done by discussion or coordination with the resource producers” (Wellcome Trust 2011).

By detecting the presence of A. pilosellae in Canada, the process for approving this candidate biocontrol agent for release in North America may potentially be expedited. Regardless of the outcome of this particular case, there is clear potential significance of both (i) the DNA barcoding of candidate biocontrol agents and (ii) making publically available standardised DNA sequences even from un-identified specimens. While DNA sequencing of candidate biological control agents is becoming more common (Gaskin et al. Reference Gaskin, Bon, Cock, Cristofaro, Biase and De Clerck-Floate2011), in large part to confirm host associations and investigate cryptic genetic variation, we advocate a protocol whereby standardised DNA regions (such as the CO1 DNA barcode) of candidate species are not only sequenced, but then routinely compared with available sequences in databases such as GenBank and BOLD. Such a protocol may reduce the amount of time required to petition candidate agents, should the species already have established but gone undocumented in the target country.