Comparative morphology of the postpharyngeal gland in the Philanthinae (Hymenoptera, Crabronidae) and the evolution of an antimicrobial brood protection mechanism

Background Hymenoptera that mass-provision their offspring have evolved elaborate antimicrobial strategies to ward off fungal infestation of the highly nutritive larval food. Females of the Afro-European Philanthus triangulum and the South American Trachypus elongatus (Crabronidae, Philanthinae) embalm their prey, paralyzed bees, with a secretion from a complex postpharyngeal gland (PPG). This coating consists of mainly unsaturated hydrocarbons and reduces water accumulation on the prey’s surface, thus rendering it unfavorable for fungal growth. Here we (1) investigated whether a North American Philanthus species also employs prey embalming and (2) assessed the occurrence and morphology of a PPG among females of the subfamily Philanthinae in order to elucidate the evolution of prey embalming as an antimicrobial strategy. Results We provide clear evidence that females of the North American Philanthus gibbosus possess large PPGs and embalm their prey. The comparative analyses of 26 species from six genera of the Philanthinae, using histological methods and 3D-reconstructions, revealed pronounced differences in gland morphology within the subfamily. A formal statistical analysis based on defined characters of the glands confirmed that while all members of the derived tribe Philanthini have large and complex PPGs, species of the two more basal tribes, Cercerini and Aphilanthopsini, possess simple and comparatively small glands. According to an ancestral state reconstruction, the complex PPG most likely evolved in the last common ancestor of the Philanthini, thus representing an autapomorphy of this tribe. Conclusion Prey embalming, as described for P. triangulum and T. elongatus, and now also for P. gibbosus, most probably requires a complex PPG. Hence, the morphology and size of the PPG may allow for inferences about the origin and distribution of the prey embalming behavior within the Philanthinae. Based on our results, we suggest that prey embalming has evolved as an antimicrobial strategy in and is restricted to the tribe Philanthini, which seems to face exceptional threats with regard to fungal infestations of their larval provisions. Electronic supplementary material The online version of this article (doi:10.1186/s12862-015-0565-0) contains supplementary material, which is available to authorized users.


Specimens and rearing conditions
Female P. gibbosus were collected on flowers in Madison (Wisconsin, USA) on August 15 th and 16 th , 2009, and transported alive to the University of Regensburg (Bavaria, Germany). The beewolves were kept in observation cages as described earlier for P. triangulum [1], but because of the smaller size of P. gibbosus, the thickness of the sand layer in the nesting compartment was reduced to about 6 mm.
The female P. gibbosus were supplied ad libitum with honey and halictid bees (Hymenoptera, Halictidae) that were collected daily in the Botanical Garden of the University of Regensburg. In order to assess whether P. gibbosus females embalm their prey with hydrocarbons (HCs) from their postpharyngeal gland (PPG), as has previously been described for P. triangulum [2,3] and T. elongatus [4], six paralyzed bees were removed from two artificial P. gibbosus brood cells (three bees each; hereafter referred to as 'provisioned bees') for chemical analysis. For comparison, nine halictid bees were collected in the field and analyzed without prior contact to P. gibbosus (hereafter referred to as 'control bees'). Additionally, heads of four P. gibbosus females that had been collected in Salt Lake City and kindly provided by Jon Seger (University of Utah) were extracted for analysis of their HC composition. The analysis of whole heads yields similar results as the analysis of dissected PPG reservoirs since the PPG contains comparatively huge amounts of HCs and their composition is virtually identical to that of the cuticle [4,5].

Molecular identification of prey species
To reliably identify the halictid bee species, we sequenced a portion of the nuclear long-wavelength rhodopsin (lwrh) and the mitochondrial cytochrome oxidase I (coxI) gene, respectively (Table S1).

Chemical analysis
Specimens were extracted for 10 minutes in approximately 1 ml of hexane. For the bee samples 2 µg of octadecane were added as an internal standard, to allow the quantification of the absolute amounts of the compounds. For each sample, the solvent was evaporated under a gentle stream of nitrogen, then 50-100 µl hexane were added, and the extract was transferred to a 200 µl GC-µ-vial (CZT, Kriftel, Germany). An aliquot of 1 µl of each sample was injected into a Agilent 6890N Series GC system coupled to a Agilent 5973 insert mass selective detector (Agilent Technologies, Böblingen, Germany). The GC was equipped with a nonpolar RH-5ms+ fused silica capillary column (30 m x 0.25 mm ID; df = 0.25 µm; Capital Analytical Ltd., Leeds, UK; temperature program: from 60°C to 300°C at 5°C/min and held for 1 min at 60°C and for 10 min at 300°C). Helium was used as the carrier gas, with a constant flow of 1 ml/min. A split/splitless injector was operated at 250°C in the splitless mode (60s). Electron impact mass spectra were recorded with an ionization voltage of 70 eV, a source temperature of 230°C, and an interface temperature of 315°C. The software MSD ChemStation for Windows was used for data acquisition. N-Alkanes were identified by the comparison of their retention times and mass spectra to those of synthetic reference substances. Linear retention indices (LRIs) for all other substances were calculated according to Van den Dool and Kratz [7] and alkenes were identified by their LRIs and mass spectra as described in Strohm et al. [5]. The structure of the unsaturated ketone nonacosen-6-one was tentatively assigned by its mass spectrum as described previously [4].

Data analysis
Besides the peaks that were identified as alkanes, alkenes, and alkadienes, we detected 29 additional substances in the extracts of halictid bees, none of which occurred in P. gibbosus samples (with the exception of one putative hexacosene, see Table S3). Since the aim of this investigation was to assess whether P. gibbosus females apply substances to the surface of their prey, no further efforts were made to identify these substances, occurring only in bee samples. As the peaks of the different isomers of the alkenes in some cases were not completely separated in the GC profile, the peak areas of all isomers of a given chain length were generally combined for further analysis. The total amounts of substances for each individual sample of provisioned and control bees were calculated using the internal standard. Since the P. gibbosus samples had been analyzed without internal standard, absolute amounts were not calculated for this group. Additionally, the total peak area was standardized to 100% and the relative amounts of substances were calculated for each individual sample from all three groups. Inspection of the chromatograms revealed considerable differences between samples of the three groups in the peak areas of the three alkenes dominating the GC profile of female P. gibbosus, namely heptacosene, nonacosene, and hentriacontene. Therefore, the proportions of each of these substances were compared between provisioned and control bees. To assess whether provisioned bees carried larger amounts of HCs and a higher proportion of unsaturated HCs, we compared the total amounts of cuticular substances, as well as the relative amounts of unsaturated hydrocarbons between provisioned bees and control bees. All relative values were arcsin-transformed prior to analysis. All statistical comparisons were conducted with t tests (using test statistics for equal or unequal variances, respectively, depending on the results of preceding Levene's tests for homogeneity of variance), using the statistics software package PAST (Version 2.08b) [8].

Comparative morphology of head glands: Coding of character states
As the result of a comprehensive examination of both semithin histological sections and 3Dreconstructions of the head glands of female Philanthinae, we defined 13 morphological characters for the comparative analysis of the PPG and MG of the 26 philanthine species under study (Table 1          (2) many hairs 13. Gland cells associated with the MG. Typically the MG of Hymenoptera is associated with class III gland cells (classified according to Noirot and Quennedey [9]), that can be identified by the occurrence of end apparatuses and conducting canals. However, in some of the investigated species there are no class III gland cells associated with the MG; however, there are cells that may be classified as class I gland cells, but these cells could not unambiguously be identified as secretory cells with the applied light-microscopic methods. Hence, we distinguished only cases with class III gland cells and those without these.
(0) the MG is not associated with class III gland cells (1) the MG is associated with class III gland cells

Additional References
Additional Tables and Figures   Table S1 Bee species included in the chemical analysis.

Group Species Blast identity Gene
Provisioned bees Table S2 Character matrix for the statistical analysis of head gland morphology of female Philanthinae.