Competition experiments.

(A) Experimental set-up of the co-culture competition experiments between ace2/ace2 and ACE2/ACE2 strains. A 50/50 mix of the two strains based either on phenotype (snowflake vs planktonic, microscopy) or on genotype (ace2 vs ACE2, qPCR) was inoculated in SDcasaWAU liquid medium. The resulting co-culture was grown for 2 days until stationary phase was reached and then diluted in new SDcasaWAU medium. For each dilution step, the proportion of the snowflake phenotype or ace2 genotype was determined (by microscopy or qPCR respectively). (B) Evolution of the percentage of snowflake entities during independent co-culture competitions between ace2/ace2 and ACE2/ACE2 strains, either cln3/cln3 (red bars) or CLN3/CLN3 (black bars) (N=4, n>250, mean ± SD, ns p>0.05, *p<0.05, ***p<0.001). The percentage of snowflakes is statistically compared between CLN3/CLN3 and cln3/cln3 co-cultures at each dilution round, using a Fisher’s exact test. (C) The proportion of ace2/ace2 genotype was monitored by qPCR in independent competitions in cln3/cln3 (red bars) and CLN3/CLN3 (black bars) backgrounds. The ACT1 locus, amplified in all the cells of the population, was used to normalize the proportion of ace2 cells in the co-culture cell population (N=3, mean ± SD, unpaired t-test, Welch correction, ns p>0.05, ***p<0.001). (D) Evolution of the percentage of snowflake entities during independent co-culture competitions between ace2/ace2 and ACE2/ACE2 strains overexpressing WHI5 (orange bars) or not (black bars) (N=3, n>235, mean ± SD, Fisher’s exact test, ns p>0.05, *p<0.05, **p<0.005, ***p<0.001).

Quiescence exit efficiency explains the results of competition experiments.

(A) Lag phase duration after cell 2-day old population refeeding with SDCasaWAU medium. The optical density at 600nm was followed after population refeeding for 4 independent experiments. The mean and the SD are indicated together with an unpaired t-test, Welch correction, ns p>0.05, ***p<0.001. (B) Representative image series of cln3/cln3 cells either ace2/ace2 or ACE2/ACE2 after refeeding of 2 days cultures on a SDCasaWAU medium containing microscope pad. Arrows point to emerging buds. Bars are 10µm. (C) Proportion of cln3/cln3 cells that have formed a new bud after refeeding on a SDCasaWAU medium containing microscope pad of either ace2/ace2 or ACE2/ACE2 2 days cultures (N=3, n>101, Fisher’s exact test, mean ± SD, ns p>0.05, *p<0.05, **p<0.005, ***p<0.001). (D) Same as C in a cln3 haploid background (N=3, n>100, mean ± SD, Fisher’s exact test, **p<0.005, ***p<0.001). (E) Percentage of snowflake entities observed in independent co-cultures of ace2/ace2 and ACE2/ACE2 strains. Cultures either went through cycles of proliferation and stationary phase as described in Fig1A (right panel), or maintained in constant exponential phase (left panel) (N>3, n>208, mean ± SD, Fisher’s exact test, ns p>0.05). (F) Same as A except that the refeeding was done after a culture of 5 days (N=3, mean ± SD, unpaired t-test, Welch correction, **p<0.005). (G) Percentage of snowflake entities observed in independent co-cultures of ace2/ace2 and ACE2/ACE2 strains diluted every two days (red and black bars) or every five days (grey and pink bars) (N≥3, n>700, mean ± SD, Fisher’s exact test, ns p>0.05, **p<0.005, ***p<0.001).

Quintuple mutant strains do not rescue the quiescence-exit default associated with cln3 and are not selected in competition experiments.

(A) Proportion of cln3/cln3 cells, either quintuple mutant or WT, that formed a new bud after refeeding of 2 days cultures on a SDCasaWAU medium containing microscope pad (N=3, n>102, mean ± SD, Fisher’s exact test, ns p>0.05, *p<0.05). (B) Proportion of cln3/cln3 cells, either quintuple mutant or ace2/ace2, that formed a new bud after refeeding of 2 days cultures on a SDCasaWAU medium containing microscope pad (N=3, n>75, mean ± SD, Fisher’s exact test, ns p>0.05, *p<0.05, **p<0.005, ***p<0.001). (C) Evolution of the percentage of snowflake entities during independent co-culture competitions between quintuple mutant and WT strains either cln3/cln3 (red bars) or CLN3/CLN3 (black bars) (N=3, n>250, mean ± SD, Fisher’s exact test).

Rescue of the quiescence-exit default associated with cln3 depends on the ace2 mutation and not on the Snowflake phenotype.

(A) A 2-day ace2/ace2 cln3/cln3 culture was sonicated and refed onto a fresh SDCasaWAU microscope pad. The proportion of cells that formed a new bud after refeeding on a SDCasaWAU medium containing microscope pad was assessed within different multicellular size ranges (N=3, n>184, mean ± SD, Fisher’s exact test). (B) 2-day cln3 cultures, either ace2 or ACE2, were sonicated and refed onto a fresh SDCasaWAU microscope pad. Within entities formed of 1 to 3 cells, the proportion of cells that formed a new bud after refeeding on a SDCasaWAU medium containing microscope pad was assessed (N=3, n>63, mean ± SD, Fisher’s exact test, ns p>0.05, *p<0.05, **p<0.005, ***p<0.001).

Rescue of the quiescence-exit default associated with cln3 is dependent on the Kss1 MAP kinase, which favors transcription of the CLN1 cyclin gene.

(A) Proportion of cln3 kss1 cells, either ace2 or ACE2, that formed a new bud after refeeding of 2-day cultures on a fresh SDCasaWAU medium containing microscope pad (N=3, n>86, mean ± SD, Fisher’s exact test, ns p>0.05). (B) Ratio of CLN1 mRNA level and ACT1 mRNA level, determined by RT-qPCR, 15 minutes after refeeding of 2 days cln3 cultures, either ace2 or ACE2 (N=3, mean ± SD, unpaired t-test, Welch correction, **p<0.005). (C) Ratio of CLN1 mRNA level and ACT1 mRNA level, determined by RT-qPCR, 15 minutes after refeeding of 2-day kss1 cln3 cultures, either ace2 or ACE2 (N=3, mean ± SD, unpaired t-test, Welch correction, ns p>0.05).

Quiescence exit efficiency and selection in competition experiments of Snowflake AMN1 strains.

(A) Proportion of cln3 cells, either AMN1 or amn1-368V, that formed a new bud after refeeding of 2 days cultures on a fresh SDCasaWAU medium containing microscope pad (N=3, n>77, mean ± SD, Fisher’s exact test, ns p>0.05, *p<0.05, **p<0.005, ***p<0.001). (B) Evolution of the percentage of snowflake entities during independent co-culture competitions between AMN1 and amn1-368V strains, either cln3/cln3 (red bars) or CLN3/CLN3 (black bars) (N=3, n>252, mean ± SD, Fisher’s exact test, ns p>0.05, **p<0.005, ***p<0.001). (C) Evolution of the percentage of snowflake entities during independent co-culture competitions between AMN1 and amn1-368V strains overexpressing WHI5 (orange bars) or not (black bars) (N=3, n>302, Fisher’s exact test, ns p>0.05, **p<0.005, ***p<0.001).

Model explaining the selection of the Snowflake phenotype as a passenger phenotype.

In strains expressing the non-functional amn1-368V allele (A), a default in quiescence exit after refeeding was observed in a cln3 mutant. In non-laboratory strains (B), the functional AMN1 allele leads to degradation of Ace2p, which can no longer repress the expression of the Kss1 Map kinase. Kss1p favors the transcription of CLN1, which would favor progression to S phase and partially suppress the default in quiescence exit associated with the cln3 mutation. This observed selective advantage (a) leads to the selection of AMN1 genotype (b), and because of the pleiotropy of Ace2p transcription factor, to the selection of the Snowflake multicellular phenotype as a passenger phenotype (c).

Strains

Plasmids

Oligonucleotides

Number of cells per Snowflake entity determined by counting the number of HTB1-CFP positive nuclei per Snowflake entity.

(A) Theoretical diagram showing the various phenomena that can explain, either alone or combined, how one subpopulation can outcompete another subpopulation in a co-culture. One subpopulation may divide more rapidly in exponential phase (a), may have a higher yield (b), may undergo less cell death during stationary phase (c) or may re-enter proliferation phase faster after refeeding (d). (B) Doubling time of populations in proliferation phase. Independent cultures (N>3) were maintained in constant proliferation phase for more than 24 hours, then the optical density at 600 nm was measured every 15 minutes for 8 hours (mean ± SD, unpaired t-test, Welch correction, ns p>0.05). (C-D) The dry biomass determined by filtration (C, N=3, mean ± SD, unpaired t-test, Welch correction, ns p>0.05) and the percentage of dead cells stained with ethylene blue (D, N=3, n>280, mean ± SD, Fisher’s exact test, ns p>0.05,) were also measured for 2 days stationary phase cultures. (E) Proportion of CLN3/CLN3 cells that have formed a new bud after refeeding on a SDCasaWAU medium containing microscope pad of either ace2/ace2 or ACE2/ACE2 2 days cultures (N=3, n>101, Fisher’s exact test, mean ± SD, ns p>0.05). (F) Proportion of CLN3 cells that have formed a new bud after refeeding on a SDCasaWAU medium containing microscope pad of either ace2 or ACE2 2 days cultures (N=3, n>91, Fisher’s exact test, mean ± SD, ns p>0.05). (G-J) Representative images of diploid ace2/ace2 snowflakes either cln3/cln3 (G) or CLN3/CLN3 (H), and haploid ace2 snowflakes either cln3 (I) or CLN3 (J). Bars are 10 µm. (K). Size distribution of snowflakes in haploid (empty circles) or diploids (plain circles) backgrounds, either CLN3 (black) or cln3 (red). Size is estimated by measuring both the longest segment of the snowflake entity and the longest segment perpendicular to the first one (N=3, n>72). We compared the length of each segment independently with an unpaired t-test with a Welch correction (error bars ± SD, unpaired t-test, Welch correction, ***p<0.001, account for both segment).

Snowflake characterization of quintuple mutant strains.

(A) Representative image of the quintuple mutant strain. Bar is 10µm. (B) Size distribution of quintuple mutant (triangles) or ace2/ace2 (circles) snowflakes. Size is estimated by measuring both the longest segment of the snowflake entity and the longest segment perpendicular to the first one. We compared the length of each segment independently with an unpaired t-test with a Welch correction (N=3, n>60, mean ± SD). (C) Proportion of snowflake entities in a population of quintuple mutant or ace2/ace2 strains (N=3, n>201, mean ± SD, Fisher’s exact test, ns p>0.05).

Effect of sonication on SF phenotype, cell death and quiescence exit.

(A-B) The effect of sonication bath on an ace2 culture was assessed by measuring the proportion of single cell entities (A, N=3, n>212) and the percentage of dead cells (B, N=3, n>207) in the population before and after the culture was sonicated. (C) Proportion of cln3 cells, either ace2 or ACE2, and either sonicated or not, that formed a new bud after refeeding of 2 days cultures on a SDCasaWAU medium containing microscope pad (N=3, n>63, Fisher’s exact test, mean ± SD).

Snowflake characterization of AMN1 strains.

(A) Representative image of the AMN1 strain. Bar is 10µm. (B) Proportion of snowflake entities in populations of AMN1 or ace2 haploid strains (N=3, n>219, mean ± SD, Fisher’s exact test, ns p>0.05). (C) Size distribution of AMN1 (diamond) or ace2 (circle) snowflakes. Size is estimated by measuring both the longest segment of the snowflake entity and the longest segment perpendicular to the first one (N=3, n>74, mean ± SD, unpaired t-test, Welch correction, ***p<0.001, account for both segment).