12Animals continuously encounter microorganisms that are essential for health or cause disease. 13 They are thus challenged to control harmful microbes while allowing acquisition of beneficial 14 microbes, a challenge that is likely especially important concerning microbes in food and in 15 animals such as social insects that exchange food among colony members. Here we show that 16 formicine ants actively swallow their antimicrobial, highly acidic poison gland secretions 17 after feeding. The ensuing creation of an acidic environment in the stomach, the crop, 18 improves individual survival in the face of pathogen contaminated food and limits disease 19 transmission during mutual food exchange. At the same time, crop acidification selectively 20 allows acquisition and colonization by known bacterial gut associates. The results of our 21 study suggest that swallowing of acidic poison gland secretions acts as a microbial filter in 22 formicine ants and indicate a potentially widespread but so far underappreciated dual role of 23 antimicrobials in host-microbe interactions. 24 25 85 5 experimental manipulation of poison gland access, and behavioural observations. In loss of 86 poison gland function experiments, we then investigate whether analogous to acidic stomachs 87 of higher vertebrates and acidic midgut regions in the fruit fly, swallowing of poison gland 88 substances can serve gut microbial control and prevent bacterial pathogen infection and 89 transmission. Finally, we explore whether swallowing of poison gland substances acts as a 90 microbial filter that is permissible to gut colonization of bacteria from the family 91 Acetobacteracea. 92 93
Results and Discussion
94To reveal whether poison gland secretions are swallowed during acidopore grooming, we 95 monitored acidity levels in the crop lumen of the Florida carpenter ant Camponotus floridanus 96 at different time points after feeding them 10% honey water (pH = 5). We found that over 97 time, the crop lumen became increasingly acidic, reaching highly acidic values 48h after 98 feeding (median pH = 2; 95% CI: 1.5-3.4), whilst renewed access to food after 48h restored 99 the pH to levels recorded after the first feeding trial ( Fig. 1a; LMM, LR-test, χ 2 = 315.18, df = 100 3, P < 0.001; Westfall corrected post-hoc comparisons: 0+4h vs. 48h+4h: P = 0.317, all other 101 comparisons: P < 0.001). This acidification was limited to the crop and did not extend to the 102 midgut (Fig. 1figure supplement 1; pH-measurements at four points along the midgut 24h 103 after access to 10% honey-water; mean ± se; midgut position 1 = 5.08 ± 0.18, midgut position 104 2 = 5.28 ± 0.17, midgut position 3 = 5.43 ± 0.16, midgut position 4 = 5.31 ± 0.19). Prevention 105 of acidopore grooming in C. floridanus ants for 24h after feeding resulted in a significantly 106 diminished acidification of the crop lumen (Fig. 1b; LMM, LR-test, χ 2 = 44.68, df = 1, P < 107 0.001), a result that was invariably obtained in a comparative survey across seven formicine 108 ant species (g...
