In preprints: quantitative decoupling between regulatory modules safeguards phenotypic robustness

As eloquently articulated by the British developmental biologist Conrad Waddington – ‘The form of a living plant or animal is continuously kept in being in spite of the fact that material is passing all the time through the system’ – living organisms are remarkably robust to adverse conditions (Waddington, 1957). The ability of biological systems to remain insensitive to genetic and environmental perturbations is known as developmental robustness or canalization. Robustness is observed ubiquitously across phyla, guaranteeing invariable phenotypic outcome (Whitacre, 2012). It is also considered to facilitate the evolution of complex dynamic systems. A resilient phenotypic trait against external variations could be adapted and selected by evolution (Kitano, 2004). Despite its known significance in biology, there is limited understanding of the mechanisms that contribute to developmental stability.

To gain insight into the underlying mechanism of developmental robustness, in this preprint Rodrigues and colleagues studied the asymmetric division of the one-cell Caenorhabditis elegans embryo (Rodrigues et al., 2023 preprint). The molecular pathways regulating this process have been very well characterized and understood (Rose and Gonczy, 2014). Following fertilization, symmetry-breaking cues polarize the embryo along the anterior-posterior axis. Two cortical domains are formed that define cell polarity. The anterior domain consists of membrane-bound PAR (partitioning) polarity complex proteins (aPAR) PAR-3, PAR-6 and atypical protein kinase C (aPKC), whereas the posterior domain encompasses polarity proteins (pPAR) PAR-1 and PAR-2. This polarity is maintained by mutual inhibition between the anterior PAR (aPAR) and posterior PAR (pPAR) proteins. Next, PAR polarity modulates the positioning of the mitotic spindle and the segregation of fate determinants that lead to successful asymmetric division.

Polarization and asymmetric division of C. elegans embryos is an ideal developmental system for understanding robustness, as these processes are highly reproducible and are invariant to genetic perturbations (Delattre and Goehring, 2021), temperature variability (Begasse et al., 2015; Neves et al., 2015) and mechanical deformations (Hench et al., 2009; Jelier et al., 2016). Moreover, embryonic development proceeds normally in heterozygous PAR mutants, indicating that the system is insensitive to variations in PAR gene dosage (Kemphues et al., 1988; Watts et al., 1996). This observation challenges the view that maintaining a balance between aPAR and pPAR proteins is crucial for proper cell division (Goehring et al., 2011; Labbé et al., 2006; Lim et al., 2021; Watts et al., 1996). In this context, Rodrigues and colleagues first investigated whether reduced PAR gene dosage corresponds to a decrease in PAR protein level. Following this, they analyzed in detail the link between PAR protein dosage and the robustness of asymmetric division. Specifically, they examined the extent to which the polarity pathway is insensitive to PAR dosage and how it compensates for the reduced level/activity of the PAR proteins.

Because heterozygous PAR mutants are viable, the authors wondered whether this viability was due to dosage compensation. Prior to this work, it was unclear whether PAR heterozygotic mutants exhibited reduced protein amounts and, if so, to what extent. They assessed whether the loss of one allele of the PAR proteins could be offset by the overexpression of the other allele. They quantified GFP levels in embryos that were (1) homozygous for GFP-tagged PAR alleles (GFP/GFP), (2) heterozygous with one GFP-tagged PAR allele and one non-tagged wild-type allele (GFP/+), and (3) heterozygous with one GFP-tagged PAR allele and the other allele either mutated or depleted by RNAi (GFP/−). Intriguingly, the GFP level in GFP/− heterozygous embryos increased only moderately compared with both GFP/GFP and GFP/+ embryos, indicating a partial compensatory regulation of PAR proteins. Given that polarization requires a balance between aPAR and pPAR proteins, the authors also examined whether the reduction of aPARs would lead to a decrease in pPAR levels or vice versa. They found no correlation between the levels of the two protein complexes. Taken together, these results showed that heterozygous PAR mutants show minimal dosage compensation, and have 50-70% less PAR protein level compared with wild-type embryos.

In their subsequent analysis, Rodrigues and colleagues examined the quantitative relationship between PAR protein levels and asymmetric cell division. They measured the cell size asymmetry and cell cycle asynchrony of the anterior AB and posterior P daughter blastomeres across varying PAR protein concentrations. The PAR protein levels were modulated by employing RNA interference (RNAi). Up to a 50% reduction in PAR protein dosage caused no substantial defect. However, a further reduction in the PAR dosage revealed significant phenotypic variations.

Similar to asymmetric division, polarization remained unaffected by the modified PAR dosages, although depleting any of the PAR complex proteins – whether anterior or posterior – led to a reduced membrane concentration of the respective PAR proteins, accompanied by ectopic localization. The overall PAR asymmetry, quantified by the PAR asymmetry index, was only modestly affected. Hence, the authors deduce that despite the PAR network's sensitivity to dosage variations, the polarity itself demonstrates robustness.

To elucidate further the robustness of asymmetry, the authors investigated how molecular pathways downstream of the PAR network react to varying PAR dosages. They focused their study on two pathways pivotal for asymmetric division: (1) spindle positioning and (2) cell fate segregation. In the late anaphase stage, the spindle is positioned asymmetrically towards the embryo's posterior, determining the division plane, and thereby creating a larger anterior AB daughter cell and a smaller posterior P blastomere (Rose and Gonczy, 2014). Cortical PAR domains establish an opposing gradient of PAR-1 and the RNA-binding proteins MEX-5 and MEX-6 along the anterior-posterior axis (Daniels et al., 2010; Griffin et al., 2011; Tenlen et al., 2008). This gradient ensures the unequal segregation of cell fate determinants and cell cycle regulators (Kipreos and Van Den Heuvel, 2019). Heterozygous aPAR and pPAR mutants did not show any defect in spindle elongation or positioning. Also, the MEX-5 and PAR-1 asymmetry remained insensitive to intermediate changes in cortical PAR protein concentration, indicating a quantitative decoupling between the PAR network and its downstream pathways, which serves as a safeguard for phenotypic outcomes (asymmetric cell size and cell fate inheritance), against variable PAR dosages.

Thereafter, Rodrigues and colleagues probed whether a similar non-linearity is observed between upstream symmetry-breaking cues and the polarity output. They examined the impact of diminished cortical actomyosin flow, a symmetry-breaking cue, on the posterior PAR-2 domain size. Attenuation of cortical flow by the knockdown of myosin regulatory light chain to velocities as low as 0.05 µm/s did not compromise PAR-2 domain size, suggesting that the polarity axis is resistant to the strength of guiding cues. However, reducing cue strength rendered the asymmetric division sensitive to PAR dosage. PAR-2 domains failed to form when both the cue strength and PAR dosage were simultaneously compromised.

In summary, this preprint highlights a non-linear relationship among the various modules necessary for asymmetric division: symmetry-breaking cues, PAR polarity, and spindle/fate asymmetry. Each module remains quantitatively insulated from the others, as long as variations remain below a certain threshold. This decoupling ensures errors in one module do not propagate to subsequent ones, thereby maintaining the robustness of the asymmetric division pathway against genetic, environmental or stochastic variations. Understanding how this quantitative decoupling is established and the degree to which this mechanism is conserved across different systems presents an intriguing avenue for future research.