Changes in interactions over ecological time scales influence single-cell growth dynamics in a metabolically coupled marine microbial community

Microbial communities thrive in almost all habitats on earth. Within these communities, cells interact through the release and uptake of metabolites. These interactions can have synergistic or antagonistic effects on individual community members. The collective metabolic activity of microbial communities leads to changes in their local environment. As the environment changes over time, the nature of the interactions between cells can change. We currently lack understanding of how such dynamic feedbacks affect the growth dynamics of individual microbes and of the community as a whole. Here we study how interactions mediated by the exchange of metabolites through the environment change over time within a simple marine microbial community. We used a microfluidic-based approach that allows us to disentangle the effect cells have on their environment from how they respond to their environment. We found that the interactions between two species—a degrader of chitin and a cross-feeder that consumes metabolic by-products—changes dynamically over time as cells modify their environment. Cells initially interact positively and then start to compete at later stages of growth. Our results demonstrate that interactions between microorganisms are not static and depend on the state of the environment, emphasizing the importance of disentangling how modifications of the environment affects species interactions. This experimental approach can shed new light on how interspecies interactions scale up to community level processes in natural environments.


Single cell growth rates depend on a cell's position in the channel
We assessed whether variation in micro-environments contributes to variation in cell growth rates. Nutrient availability in the growth channels is determined by the balance of nutrients diffusing into the growth channel from the flow channel and the uptake of these nutrients by cells. When nutrient concentrations in the flow channel are low and uptake rates are high, microscale gradients can be formed. These gradients can affect single cell growth rates across microfluidic channels. In natural environments these gradients can be the result of diffusion from a particulate nutrient source. Cells that are closer to the nutrient source will have immediate access to nutrients while cells that are further away (i.e. in a biofilm surrounding the nutrient core) will be more affected by diffusion and gradients.
In order to investigate which role gradients play in our experiments, we studied the correlation between single cell growth rates and the relative position of individual cells in the channels. We found that for the degrader single cell growth rates were generally higher when cells were closer to the main channel ( Figure S5C, S5D). This effect is constant over the period of the experiment ( Figure S6). The cross-feeder's growth rate is not affected by the position in the microfluidic device. This indicates that degrader and cross-feeder are generally differentially affected by growth in the microfluidic device, suggesting that the nutrients that the degrader needs are likely available at low concentrations. The degrader when grown on mono-culture (dark red) shows a correlation between single cell growth rate and a cell's position in the microfluidic device (R = 0.46, p < 0.001).     Comparison of average growth rate of degrader cells between the two main growth phases for different position in the growth channel. Single cell growth rates for the lower part of the microfluidic channel (relative position in channel 0-100) were averaged for each time point ( Figure S2) and binned into two 20h time windows. During the first 20h the single cell growth rates were significantly higher when the degrader experienced a co-culture environment (red) compared to a mono-culture condition (yellow).
Analysis using a mixed effect model (with the fixed effect culture-type and the random effect day of the experiment) revealed significant higher growth rates for the co-culture condition; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.1, std.Error = 0.03, p << 0.001). Growth during the secondary growth phase is significantly higher in mono-culture; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.37, std.Error = 0.03, p << 0.001). and binned into two 20h time windows. During the first 20h the single cell growth rates were significantly higher when the degrader experienced a co-culture environment (red) compared to a mono-culture condition (yellow). Analysis using a mixed effect model (with the fixed effect culture-type and the random effect day of the experiment) revealed significant higher growth rates for the co-culture condition; (N(Coculture) = 956; N(Monocul ture) = 956, growth rate difference = 0.06, std.Error = 0.02, p = 0.002). Growth during the secondary growth phase is significantly higher in mono-culture; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.45, std.Error = 0.02, p << 0.001).
Figure S14: Single cell growth rate differences along microfluidic growth channel do not influence overall differences in growth dynamics.
Comparison of average growth rate of cross-feeder cells between the two main growth phases for different position in the growth channel. Single cell growth rates for the lower part of the microfluidic channel (relative position in channel 0-100) were averaged for each time point ( Figure S2) and binned into two 20h time windows. During the first 20h the single cell growth rates were significantly higher when the cross-feeder experienced a co-culture environment (red) compared to a mono-culture condition (green). Analysis using a mixed effect model (with the fixed effect culture-type and the random effect day of the experiment) revealed significant higher growth rates for the mono-culture condition; ( N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.04, std.Error = 0.007, p << 0.001). Growth during the secondary growth phase is significantly higher in mono-culture; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.67, std.Error = 0.01, p << 0.001).
Figure S14: Single cell growth rate differences along microfluidic growth channel do not influence overall differences in growth dynamics.
Comparison of average growth rate of cross-feeder cells between the two main growth phases for different position in the growth channel. Single cell growth rates for the lower part of the microfluidic channel i.e. close to the outlet (relative position in channel 100-200) were averaged for each time point ( Figure S2) and binned into two 20h time windows. During the first 20h the single cell growth rates were significantly higher when the cross-feeder experienced a co-culture environment (red) compared to a mono-culture condition (green). Analysis using a mixed effect model (with the fixed effect culture-type and the random effect day of the experiment) revealed significant higher growth rates for the monoculture condition; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.03, std.Error = 0.009, p < 0.001). Growth during the secondary growth phase is significantly higher in mono-culture; (N(Co-cul ture) = 956; N(Monocul ture) = 956, growth rate difference = 0.63, std.Error = 0.01, p << 0.001).