Classifying interactions in a synthetic bacterial community is hindered by inhibitory growth medium

Abstract

Predicting the fate of a microbial community and its member species relies on understanding the nature of their interactions. However, designing simple assays that distinguish between interaction types can be challenging. Here, we performed spent media assays based on the predictions of a mathematical model to decipher the interactions between four bacterial species: Agrobacterium tumefaciens (At), Comamonas testosteroni (Ct), Microbacterium saperdae (Ms) and Ochrobactrum anthropi (Oa). While most experimental results matched model predictions, the behavior of Ct did not: its lag phase was reduced in the pure spent media of At and Ms, but prolonged again when we replenished with our growth medium. Further experiments showed that the growth medium actually delayed the growth of Ct, leading us to suspect that At and Ms could alleviate this inhibitory effect. There was, however, no evidence supporting such "cross-detoxification" and instead, we identified metabolites secreted by At and Ms that were then consumed or "cross-fed" by Ct, shortening its lag phase. Our results highlight that even simple, defined growth media can have inhibitory effects on some species and that such negative effects need to be included in our models. Based on this, we present new guidelines to correctly distinguish between different interaction types, such as cross-detoxification and cross-feeding.

Classifying Interactions in a Synthetic Bacterial Community Is Hindered by Inhibitory Growth Medium

“…We reran our original experiment in a plate reader to obtain higher time resolution, and indeed, we found a significant difference in the final yields of C. testosteroni in MM and the pure SM of A. tumefaciens ( P < 0.01) and M. saperdae ( P < 0.001) as well as a small “bump” at the beginning of the SM growth curves ( Fig. 4F , arrows) (complete statistical results on the final yield and the length of the lag phase are available in reference 32 ). These two features, obtained by high-resolution growth curve measurements until stationary phase, make it possible to distinguish between cross-feeding and cross-detoxification without the need for further molecular analyses.…”

“…As LC-MS yielded only the relative abundances of each compound, we added several concentrations to C. testosteroni growing in MM. We found that a range of concentrations shortens the lag phase of C. testosteroni significantly (oxoglutarate, 0.39 mM to 12.5 mM; proline, ≥0.39 mM; hypoxanthine, ≥0.62 mM [complete statistical results are available in reference 32 ]). At high-enough concentrations, all three metabolites also increased the final yield of C. testosteroni , suggesting that they act as carbon sources (oxoglutarate, 0.09 mM to 12.5 mM; proline, ≥0.78 mM; hypoxanthine, ≥1.25 mM [complete statistical results are available in reference 32 ]).…”

“…Each culture was then washed 2 times in PBS (phosphate-buffered saline) (centrifugation for 15 min at 4,000 rpm at room temperature), and the final bacterial pellets were resuspended in adequate medium so that the initial OD 600 would be 0.1. The compositions of the different media used are available elsewhere ( 32 ).…”

“…We grew the 4 species in monocultures under these 5 conditions ( V = 4 mL) over 60 h, measured the OD 600 over time (using an Ultrospec 10 cell density meter; Biochrom), and performed CFU counts before the initial incubation, at 24 h and 48 h. We calculated the area under the OD 600 growth curve (AUC) and used this value as a proxy for growth (DescTools [ 62 ]; R version 4.1.2). We repeated these SM assays in 96-well plates ( V = 200 μL) to increase the resolution of our data and measured the OD 600 every 10 min for 72 h using a microplate reader (BioTek Synergy H1 at 28°C with continuous double-orbital shaking) (see reference 32 for the full data set).…”

“…We generated SM for each species as described above (see the section on spent medium assays, above) (total incubation time of ~89 h) according to the specific protocols of the following kits to determine which concentration of SM to test given the theoretical concentrations of the tested compounds in fresh MM: a glucose (HK) assay kit (catalog number GAHK-20; Sigma), a citric acid kit (catalog number K-CITR; Megazyme), a phosphate assay kit (colorimetric) (catalog number ab65622; Abcam), a sodium assay kit (colorimetric) (catalog number MAK247; Sigma), and a potassium assay kit (fluorometric) (catalog number ab252904; Abcam). Osmolarity was measured in each SM sample using an osmometer (Osmomat 030; Gonotec) (the full data set is available in reference 32 ).…”