Protection of a defensive symbiont does not constrain the composition of the multifunctional hydrocarbon profile in digger wasps

Hydrocarbons (HCs) fulfil indispensable functions in insects, protecting against desiccation and serving chemical communication. However, the link between composition and function, and the selection pressures shaping HC profiles remain poorly understood. Beewolf digger wasps (Hymenoptera: Crabronidae) use an antennal gland secretion rich in linear unsaturated HCs to form a hydrophobic barrier around their defensive bacterial symbiont, protecting it from brood cell fumigation by toxic egg-produced nitric oxide (NO). Virtually identical HC compositions mediate desiccation protection and prey preservation from moulding in underground beewolf brood cells. It is unknown whether this composition presents an optimized adaptation to all functions, or a compromise due to conflicting selection pressures. Here, we reconstitute the NO barrier with single and binary combinations of synthetic linear saturated and unsaturated HCs, corresponding to HCs found in beewolves. The results show that pure alkanes as well as 3 : 1 mixtures of alkanes and alkenes resembling the composition of beewolf HCs form efficient protective barriers against NO, indicating that protection can be achieved by different mixtures of HCs. Since in vitro assays with symbiont cultures from different beewolf hosts indicate widespread NO sensitivity, HC-mediated protection from NO is likely important across Philanthini wasps. We conclude that HC-mediated protection of the symbiont from NO does not exert a conflicting selection pressure on the multifunctional HC profile of beewolves.


Introduction
Cuticular hydrocarbons (CHCs) protect insects from desiccation, and have often evolved secondary functions, e.g.intra-and interspecific communication in social and solitary insects [1][2][3].Constituents of insect CHC profiles include n-alkanes, methyl-branched alkanes, alkenes and alkadienes of different chain lengths [2], forming complex compositions with up to more than 100 different components [1].The chemical composition influences the function of the profile [3].One profile frequently fulfills multiple functions, e.g.cuticle lubrication [4], prey protection from pathogens [5,6] and enhancing tarsal adhesion [7], posing potentially conflicting requirements on the CHC composition [3].However, due to its complexity, it remains poorly understood how composition influences function, and how natural selection shapes composition considering the functional constraints [3].
The nearly identical HC profiles of the cuticle, the PPG and the AGS of P. triangulum are characterized by an approximately 3 : 1 alkene-alkane ratio, with compounds ranging from C21 to C31 in chain length [17][18][19][20].Tricosane (C23) represents the most abundant alkane [17][18][19][20], and either pentacosene (C25 : 1) or heptacosene (C27 : 1) constitute the most abundant alkene [17,19,20].Together with tricosane, C25 : 1 or C27 : 1 account for 51-92% of HCs [17][18][19][20].Thus, desiccation protection, prey preservation and symbiont protection are realized by virtually the same profile (electronic supplementary material, table S1, figure S1) [17][18][19][20].However, it remains unclear whether the alkene-rich HC profile presents an adaptation to all three functions, or a compromise arising from conflicting selection pressures.Here we tested single and binary mixtures of synthetic linear saturated and unsaturated HCs, corresponding to those found in beewolves, for their effectiveness in blocking NO.We demonstrate that a range of individual HCs, and mixtures of alkenes and alkanes resembling beewolf HC extracts, are effective NO barriers in vitro.Additionally, we show that symbiont strains from multiple different host species are susceptible to NO in vitro in the absence of HCs, and that CHC profiles across different beewolf species have similar compositions, indicating a widespread HC-based protection of defensive symbionts across Philanthini.While our findings support the important function of HCs in the AGS, we argue that AGS-mediated symbiont protection does not exert a conflicting selection pressure on the multifunctional HC profile and thus does not constrain its composition.

Results and discussion
We assessed NO sensitivity of five 'S.philanthi' biovariations from all host genera [21] and six free-living Streptomyces species (electronic supplementary material, table S2) to five NO concentrations.Survival was assessed as a trinary response (no growth, growing slower than control (without NO), growing as quickly as control).Most of the symbiont strains were already affected by low NO concentrations, with growth completely ceasing at 1% NO.By contrast, most free-living strains were unaffected at concentrations below 1% and still grew at 1% NO.In the statistical analysis, our final model retained 'bacterial category/strain' and 'NO concentration' as independent variables (electronic supplementary material, methods).Symbionts were significantly more sensitive to NO than freeliving Streptomyces (ANOVA factor 'bacterial category', χ 2 = 103.5,d.f.= 2, p < 2.2 × 10 −16 , figure 1), and strains varied in their NO sensitivity (ANOVA factor 'bacterial category/ strain', χ 2 = 155.4,d.f.= 40, p = 1.618 × 10 −15 ).Expectedly, the impact of NO on bacterial growth increased with concentration (ANOVA factor 'NO concentration', χ 2 = 200.3,d.f.= 2, p < 2.2 × 10 −16 ).Given the symbionts' sensitivity towards NO, we hypothesize that other beewolf hosts beyond P. triangulum may use their HCs in the AGS to protect their symbionts from NO fumigation.
To examine the link between HC composition and NO barrier function, we tested synthetic binary alkene-alkane  S2 for strain designations.
Our experiments revealed that the hydrophobic NO barrier can be reconstituted by single and binary HC combinations and is not specific to certain HCs, implying a general effect.As previously observed in phase behaviour [22], binary combinations did not behave as predicted from individual HCs: While ineffective individually, C25 : 1 enhanced the effect of C27, but not of C23.C25 : 1 did not provide a barrier to NO, but C23 : 1-covered 'S.philanthi' survived an otherwise lethal NO exposure [8].This may be explained by the difference in the HC amount applied in both experiments resulting from the need to apply the HC without a harmful solvent in the previous study.Interestingly, C25 : 1 + C27 protected better from NO than C25 : 1 + C23, although the latter more closely resembles P. triangulum CHC extracts [17].
Assuming a common site of production and/or a shared pathway for HC biosynthesis for the AGS, PPG, and cuticle, the HC composition is likely shaped by selection acting simultaneously on symbiont protection (AGS), prey preservation (PPG), and desiccation resistance (cuticle).The more complex profile of the AGS (as well as the cuticle and the PPG)-as opposed to a simpler mixture of C25 : 1 and C27-might be explained by the wider melting range of a more complex composition ensuring adequate viscosity or establishing a biphasic secretion under varying environmental conditions [23].Furthermore, less abundant HCs may serve a role in AGS localization by the larva or as nutrients for 'S.philanthi' [17,[24][25][26].Alternatively, unspecific enzymes in the HC biosynthesis may produce a homologous HC series as a byproduct, without the minor HCs being selectively favoured [3].
Their equally alkene-and alkene/alkadiene-rich HC profiles may qualify other beewolves to provide a hydrophobic NO barrier to protect their symbionts if they fumigate their brood cells.The reported NO fumigation in P. triangulum, P. gibbosus and P. basilaris, and the ecology shared among beewolves renders a widespread NO fumigation across Philanthini wasps plausible.
Apart from blocking NO, similar HC profiles mediate prey preservation [5,6,12] and protect P. triangulum from desiccation.Previous studies indicate a selection for alkene-rich profiles in Philanthinae digger wasps, including beewolves (Philanthini) and part of the Cercerini, to efficiently preserve decay-prone Hymenopteran prey (e.g.Apidae and Halictidae) [29].By contrast, basal Cercerini providing unembalmed Coleoptera possess diversifying HC profiles, with often lower amounts of alkenes [29].Furthermore, Aphilanthopini are unlikely to embalm their ant prey, which is presumably less susceptible to microbial threats [28].Thus, prey preservation likely evolved independently in the derived Hymenopterahunting Cercerini and in the ancestor of the Philanthini, the latter probably coinciding with the acquisition of defensive symbionts about 68 mya [30].Although additional evidence is needed, we speculate that the evolutionary origin of NO fumigation and the HC-mediated protection of the symbionts may have coincided with the origins of symbiosis and prey embalming (electronic supplementary material, figure S6).
Prey preservation likely selects for high proportions of alkenes [29], and our experiments indicate their suitability for NO protection.For desiccation protection, the large qualitative variety of CHC in insects [3] suggests that this function can be realized by very different compositions, provided they form a biphasic layer [23].Thus, HC-mediated protection of the symbionts does not appear to impose a conflicting selection pressure on the composition of the beewolf AGS that could otherwise compromise the efficiency of the same HC profile for desiccation resistance and prey preservation.

Methods (a) Bacterial cultivation
Bacterial strains were cultured in a 1 : 1 mixture of Sf-900 II SFM medium (Gibco, Thermo Fisher Scientific, Germany) and Grace's insect medium (electronic supplementary material, methods; Sigma-Aldrich, Germany) at 30°C in 24-well plates.

(b) Comparative cultivation assay
We assessed the NO sensitivity of five symbiont strains representing all three host genera and different geographic origins [21], and six free-living Streptomyces strains obtained from DSMZ (Braunschweig, Germany).After NO exposure (electronic supplementary material, methods), we analysed the trinary growth observation (no growth, growth observed later than in the control, growth observed at the same time as in the control; n = 2-4) using a multinominal regression model.Growth was examined as a function of NO concentration and bacterial category (symbiont versus free-living Streptomyces), with 'strain' as a nested factor within bacterial category.Starting from the full factorial model, we used a step-wise reduction of model complexity to select the best-fitting model.Statistical analyses were performed in R i386 4.1.2using the 'nnet' [31] and 'car' [32] packages.

(c) Extraction and quantification of beewolf cuticular hydrocarbons
P. triangulum females collected in Berlin, Germany, were reared in observation cages [33].We assessed CHC extracts from 19 females regarding their efficacy as an NO barrier.The females' antennae were removed.One female per extract was submerged in 1 ml hexane.After a 10 min extraction under stirring at RT, the female was removed, and hexane was evaporated under argon flow.CHCs were re-dissolved in 100 µl hexane.A 95 µl aliquot of each extract was evaporated under argon flow and stored at −20°C.The remaining 5 µl were used for GC-MS (electronic supplementary material, methods).We characterized the CHC composition of P. histrio using a single female from a collection near Knysna, Western Cape Province, South Africa, in 2005.
After removing the head, the thorax and abdomen were extracted for 30 min in hexane.The extract was subjected to GC-MS (electronic supplementary material, methods).

(d) Cuticular hydrocarbon experiments
We purchased HCs from Sigma Aldrich, Germany, and Cayman Chemical, Michigan, USA.C25 : 1 was combined with C23 or C27 in a 3 : 1 ratio to mimic the alkene-alkane ratio found in beewolves.We transferred 10 µl of hexane containing 100 µg, 50 µg or 10 µg of each treatment (N = 7-8 each), on top of 40 µl NO indicator solution (electronic supplementary material, methods) in tubes (diameter = 3 mm; Biozym, Germany), respectively.As a positive control, we transferred each beewolf CHC extract in the same way.Indicator solutions treated with 10 µl hexane served as negative controls (N = 8).The applied HCs form a distinct layer on top of the indicator solution due to their hydrophobicity.After NO exposure (electronic supplementary material, methods), the content of the tubes was centrifuged in 0.5 ml tubes at maximum speed for 30 s.We measured the OD 540 of 20 µl of the supernatant in a 384 well plate (VarioSkan Lux, Thermo Scientific, Germany).We performed OD 540 comparisons across all 100 µg treatments and both controls, and within treatments, using a one-way ANOVA.The correlation between the amount of beewolf CHCs and the OD 540 was analysed using Spearman's rank correlation.Statistical analyses were conducted in R (V4.15).

Figure 2 .
Figure2.Reconstitution of the nitric oxide (NO) barrier effect with single and binary combinations of synthetic hydrocarbons (HCs).A total of 100 µg of HCs was applied for all treatments.Effectivity of HCs in blocking NO was measured as the change in coloration in an NO indicator solution, with higher OD 540 values indicating stronger oxidation and thus less protection against NO by the HC layer (see representative images on the top; for all images see electronic supplementary material, figureS4).Beewolf CHC extracts served as positive, and hexane as negative control.All HCs and the alkene-alkane ratio of binary combinations are found in P. triangulum AGS and HC extracts (electronic supplementary material, table S1, figureS1).Letters indicate significant differences (Tukey's HSD, n = 8, p < 0.05).