Animals continuously encounter microorganisms that are essential for health or cause disease. They are thus challenged to control harmful microbes while allowing acquisition of beneficial microbes. This challenge is likely especially important for social insects with respect to microbes in food, as they often store food and exchange food among colony members. Here we show that formicine ants actively swallow their antimicrobial, highly acidic poison gland secretion. The ensuing acidic environment in the stomach, the crop, can limit the establishment of pathogenic and opportunistic microbes ingested with food and improve survival of ants when faced with pathogen contaminated food. At the same time, crop acidity selectively allows acquisition and colonization by Acetobacteraceae, known bacterial gut associates of formicine ants. This suggests that swallowing of the poison in formicine ants acts as a microbial filter and that antimicrobials have a potentially widespread but so far underappreciated dual role in host-microbe interactions.

Data Fig.1a: Source data on pH of crop lumens at 4h, 24h and 48h after feeding C. floridanus ants 10% honey water at 0h and at 4h after re-feeding ants at 48h.

Data Fig.1b: Source data on pH of crop lumens in C. floridanus ants that were either prevented to ingest formic acid containing poison gland secretions (FA-) or not (FA+) for 24h after feeding.

Data Fig.1c: Source data on pH of crop lumens 24h after feeding in seven formicine ant species that were either prevented to ingest formic acid containing poison gland secretions (FA-) or not (FA+).

Data Fig.1-supplementary figure 1: Source Data on the baseline acidity of C. floridanus (major and minor worker caste) under satiated and starved conditions.

Data Fig.1-supplementary figure 2: Source data on the frequency of acidopore grooming in C. floridanus ants within 30 min. after fluid ingestion (water or 10% honey water) compared to ants that did not receive any fluid (unfed).

Data Fig.1-supplementary figure 3: Source data on pH-measurements 24h after access to 10% honey-water in the crop and directly after the proventriculus at four points along the midgut of C. floridanus ants.

Data Fig.2a,b: Source data on the number and the change in the number of colony forming units (CFUs) relative to 0h in the crop in the crop (a) and midgut (b) part of the digestive tract of C. floridanus ants at 0h, 0.5h, 4h, 24h, and 48h after feeding ants 10% honey water contaminated with Serratia marcescens.

Data Fig.2-supplementary figure 1: Source data on the food passage of florescent particles through the digestive tract (crop, midgut, hindgut) of C. floridanus minor (a) and major (b) worker ants.

Data Fig.2-supplementary figure 2a,b: Source data on the number and the change in the number of colony forming units (CFUs) in the crop (a) and midgut (b) part of the digestive tract of C. floridanus ants relative to 0h in the crop at 0h, 0.5h, 4h, 24h, and 48h after feeding ants 10% honey water contaminated with Escherichia coli.

Data Fig.2-supplementary figure 3: Source data on the number and the change in the number of CFUs relative to pH 5 after incubation of Serratia marcescens in 10% honey water (pH = 5) or in 10% honey water acidified with commercial formic acid to a pH of 4, 3 or 2 for 2h.

Data Fig.3: Source data on the survival of individual C. floridanus ants that were either prevented to ingest formic acid containing poison gland secretions (FA-) or not (FA+) after feeding on either honey water contaminated with Serratia marcescens (Serratia+) or non-contaminated honey water (Serratia-).

Data Fig.4: Source data on the survival of donor C. floridanus ants that were directly fed with pathogen contaminated food and were either prevented to ingest formic acid containing poison gland secretions (FA-) or not (FA+) and survival of receiver ants that received pathogen contaminated food only through trophallaxis with donor ants and were always prevented to ingest formic acid containing poison gland secretions (FA-).

Data Fig.4-supplementary figure 1: Total duration of trophallaxis events within 30 min. of the first bout of food exchange between donor-receiver ant-pairs of C. floridanus ants. Donor ants in both pairs were directly fed with Serratia marcescens contaminated 10% honey water and were either prevented to ingest formic acid containing poison gland secretions (FA-) or not (FA+), while receiver ants received pathogen contaminated food only through trophallaxis with the respective donor ants and were always prevented to ingest formic acid containing poison gland secretions (FA-).

Data Fig.5a,b: Source data on the number and the change in the number of colony forming units (CFUs) relative to 0h in the crop in the crop (a) and midgut (b) part of the digestive tract of C. floridanus ants at 0h, 0.5h, 4h, 24h, and 48h after feeding ants 10% honey water contaminated with Asaia sp.

Data Fig.5-supplementary figure 1: Source data on the number and the change in the number of CFUs relative to pH 5 after incubation of Asaia sp. in 10% honey water (pH = 5) or in 10% honey water acidified with commercial formic acid to a pH of 4, 3 or 2 for 2h.

Data Code and Script file: File containing all code required to reproduce the analyses and figures in R version 3.6.1.