Developmental constraints enforce altruism and avert the tragedy of the commons in a social microbe

Significance Organisms often generate benefits shared among their whole group, but such cooperation is vulnerable to collapse if individuals can instead benefit by exploiting the cooperation of others. While relatedness can promote cooperation, many species lack reliable mechanisms to ensure high relatedness. They are therefore vulnerable to a breakdown of cooperation unless they are able to enforce cooperation. We test this idea through experimental manipulation of group composition in a social microbe. We find that groups avert the expected collapse in cooperation at low relatedness due to inadvertent enforcement of cooperation by a mechanism that prevents errors in multicellular development. Our findings explain how mechanisms that promote cooperation can arise as by-products of natural selection acting on traits in other contexts.


Robustness to model assumptions
We evaluate the robustness of the main model predictions to non-linearity of benefits using two general shapes of non-linear benefit functions: diminishing and accelerating returns (see Figure S4). For each, we derive a new function for group benefits and solve the optimal investment. The non-linear equations for group benefits are as follows: diminishing returns = 1 + [1 − (1 − ) 1.3 ] and accelerating returns = 1 + 1.3 , where is the investment of the group. Importantly, these different functions do not alter the qualitative pattern of investment by strains, but rather, they shift the expected level of investment above or below that expected from the linear function.  Figure S1. Representative images of clonal and chimeric fruiting bodies and slugs containing RFPexpressing and mock-transformed control cells. Each row contains images of groups containing the strain listed in the first column (NC52.3, NC60.1,NC63.2, NC80.1 and NC99.1) either as a clonal group or as the low relatedness labelled strain in a chimeric mixture. For the first four strains (NC52.3, NC60.1,NC63.2, and NC80.1) the low relatedness mixtures contain RFP-expressing cells of that strain at a frequency of 10% and mock-transformed cells of the strain NC99.1 at a frequency of 90%. For the row containing NC99.1, the low relatedness mixtures contain RFP-expressing NC99.1 at a frequency of 10% and mock-transformed cells from NC52.3 at a frequency of 90%. For fruiting bodies, the stalks of low relatedness mixtures show lower levels of fluorescence than the clonal mixtures. For slugs, the prespore region (anterior quarter) of low relatedness slugs shows lower fluorescence (relative to the posterior three-quarters) than the clonal mixtures (which show an even distribution of fluorescently labelled cells throughout). Figure S2. Images of aggregations used in the smFISH experiment. All images were taken at 14.5 hours into development. Each image labelled with a single strain ID represents a plate with clonal aggregations composed of that strain. The 10-way chimera is composed of a mix of these ten strains, each in equal proportion. We also include an image of a two-way chimera composed of one of the strain pairs for comparison but did not use pairwise mixes in the smFISH experiment. Figure S3. The shape of frequency dependent error. The different lines represent different exponents of the error function, with a higher exponent corresponding to a strong degree of frequency dependence in error. Each line is defined by the equation 4 t [ri(1 − ri)] t , where the lines corresponds to the value of t and the x-axis to the values of ri. Figure S4. The impact of non-linearity on patterns of strategic investment. A) the three lines show a linear relationship between investment and benefits from public goods (which is assumed in the Collective Investment Game) as well as diminishing and accelerating benefits. B) expected patterns of strategic investment as a function of relatedness for linear, diminishing, and accelerating benefits.

Slugs
Dataset S1 (separate file) Investment data for three strain group. This file contains data from groups composed of three strains mixed at varying proportions and data from clonal development for the same set of strains. The first sheet contains the data from chimeric mixes: the strain IDs and frequencies in the mix, the observed number of spores for the chimera, the number of spores produced by each strain in clonal development, the expected number of spores (based on the weighted average of the spore production by the constituent strains when clonal), and the inferred level of collective investment. The second sheet contains the data from clonal development for the same set of strains. The column of means are values averaged over technical replicates.
Dataset S2 (separate file) Investment data for N strain groups. This file contains data from groups composed of N strains, with each at a proportion of 1/N, and data from clonal development for the same set of strains. The first sheet contains the data from chimeric mixes: the strain IDs for each mix, the observed number of spores for the chimera, the number of spores produced by each strain in clonal development, the expected number of spores (based on the weighted average of the spore production by the constituent strains when clonal), and the inferred level of collective investment. The second sheet contains the data from clonal development for the same set of strains. The column of means are values averaged over technical replicates.
Dataset S3 (separate file) Fruiting body collapse data. This file contains data from groups composed of N strains, with each at a proportion of 1/N. Each row represents data from one plate containing N strains, with the set of strains indicated in the strain compositions columns: the total number of fruiting bodies on the plate, the number of collapsed fruiting bodies, and the percent collapsed.
Dataset S4 (separate file) Investment data for two and 20 strain groups. The file contains measurements of stalk investment for strains at a frequency of 5% in two-strain mixes and for chimeras composed of a set of 20 strains mixed at equal frequencies.
Dataset S5 (separate file) smFISH data. The file contains data on the dot counts from the smFISH experiment in clonal and chimeric mixtures for a set of natural strains and clonal data from the lab strain AX4. The first sheet contains data from clonal development: the experiment ID (1 or 2), the imagine number, the block ID, dot counts for ecmA and pspA, these values after censoring for low counts, the rescaled value of pspA, and the pspA index. The second sheet contains this same information for the chimeric mixes. The third sheet contains the same data for clonal development in AX4.
Dataset S6 (separate file) Fruiting body images and stalk fluorescence measures. The file contains images of fruiting bodies under fluorescence for RFP and associated measurements of stall fluorescence. Images are arranged in rows corresponding to the RFP labelled strain at a frequency of 0.1 and the columns the mock-transformed cells at a frequency of 0.9. Below each image there is a set of three repeats of the measurement of the total stalk fluorescence and three separate measures of fluorescence in areas without RFP expressing cells. The final column in each set has the means, with the difference in means below. These measures are summarized on the sheet with the processed data, which also includes an estimate of the relative maturation level of the fruiting body.
Dataset S7 (separate file) Slug images and prestalk_prespore fluorescence measures. The file contains images of slugs under fluorescence for RFP and associated measurements of prestalk and prespore fluorescence. Images are arranged in rows corresponding to the RFP labelled strain at a frequency of 0.1 and the columns the mock-transformed cells at a frequency of 0.9. Below each image there is a set of three repeats of the measurement of the fluorescence of the prespore and prestalk regions. The final column in each set has the means, with the difference in means below. These measures are summarized on the sheet with the processed data, which also includes the proportional fluorescence of the prestalk region.