A functionally divergent intrinsically disordered region underlying the conservation of stochastic signaling

Stochastic signaling dynamics expand living cells’ information processing capabilities. An increasing number of studies report that regulators encode information in their pulsatile dynamics. The evolutionary mechanisms that lead to complex signaling dynamics remain uncharacterized, perhaps because key interactions of signaling proteins are encoded in intrinsically disordered regions (IDRs), whose evolution is difficult to analyze. Here we focused on the IDR that controls the stochastic pulsing dynamics of Crz1, a transcription factor in fungi downstream of the widely conserved calcium signaling pathway. We find that Crz1 IDRs from anciently diverged fungi can all respond transiently to calcium stress; however, only Crz1 IDRs from the Saccharomyces clade support pulsatility, encode extra information, and rescue fitness in competition assays, while the Crz1 IDRs from distantly related fungi do none of the three. On the other hand, we find that Crz1 pulsing is conserved in the distantly related fungi, consistent with the evolutionary model of stabilizing selection on the signaling phenotype. Further, we show that a calcineurin docking site in a specific part of the IDRs appears to be sufficient for pulsing and show evidence for a beneficial increase in the relative calcineurin affinity of this docking site. We propose that evolutionary flexibility of functionally divergent IDRs underlies the conservation of stochastic signaling by stabilizing selection.


25
One of the most remarkable features of living cells is their ability to transmit and process 26 information about their surroundings. It is now appreciated that the dynamics of molecules 27 connected in regulatory networks and signaling pathways underlie many of these capabilities (1-28 3). But how do peptide sequences underlying this ability evolve? Relative to enzymatic functions 29 whose evolution has been studied for decades (4-7), research on the evolution of cellular  Here we consider the evolution of the pulsatile dynamics of Crz1, a transcription factor in 42 budding yeast that responds to rapid fluctuations of cytosolic calcium concentration (which we 43 refer to as calcium bursts (19)). Pulsatile dynamics are steady state stochastic fluctuations that 44 phenotype over long evolutionary time. The conservation of phenotype but lack of conservation 68 of IDR functions indicates that, within the disordered region, evolutionary changes in some 69 elements needed for complex signaling dynamics have compensated for evolutionary changes in 70 others. This pattern of compensatory evolution in the context of preserved function is a hallmark 71 of stabilizing selection (42). By comparing IDR sequences of Crz1 orthologues, we infer that one 72 of these elements is the calcineurin docking site, PxIxIT, which increased binding strength 73 during evolution. Remarkably, by experimentally increasing the PxIxIT strength in a distantly 74 related IDR to the Saccharomyces PxIxIT strength (via three point-mutations), we can rescue 75 pulsing and improve fitness. Our study demonstrates that stochastic pulsatility is beneficial and 76 that a position-dependent molecular feature in the IDR plays a role in rewiring the molecular 77 basis of the stochastic signaling pathway, even though the phenotype is preserved.

79
Evolutionary changes in Crz1 IDRs are associated with changes in Crz1 pulsing 80 We first sought to confirm that the IDR of Crz1 was responsible for pulsing. To do so, we 81 designed a passive reporter system in S. cerevisiae that expresses an IDR tagged with GFP. We 82 found that the S. cerevisiae disordered region alone showed pulsing with similar dynamics as the 83 endogenous protein, although the expression level of the protein was lower (Supplementary 84 Figure 1). Therefore, we fused a defective Crz1 DNA-binding domain (43) tagged with GFP to 85 the disordered regions (denoted as Sc-reporter) and found nearly endogenous dynamics and 86 expression levels, indicating that the disordered region is sufficient for the pulsing dynamics but 87 that the DNA binding domain is needed for protein stability (see methods for more details). 88 Since the calcium/calcineurin signaling pathway is highly conserved (26), we predicted that 89 functional elements within the disordered regions would be conserved over evolution if the 90 dynamics of Crz1 are important for signaling function (25). Consistent with this, some functional 91 elements important for the control of subcellular localization, such as the nuclear localization 92 signal (NLS), nuclear export signal (NES), and the calcineurin docking site PxIxIT (37,38) 93 (summarized in Figure 1B), are found in orthologous Crz1 sequences. On the other hand, overall, 94 the IDRs of Crz1 are highly diverged (little sequence similarity is detected in sequence 95 alignments, Figure 1C), which leads to an opposite prediction that the dynamics of Crz1 96 orthologues would diverge as has been found for p53 (9). To quantify Crz1 dynamics in response 97 to the upstream calcium signaling pathway, alongside the GFP-tagged Crz1 IDR reporter 98 (denoted as "pulsing reporter"), we expressed a calcium sensor GCaMP3 (44) in a "double-99 reporter strain" (see methods for more details). 100 We first confirmed that, as expected based on previous reports for S. cerevisiae, C.  Figure 2B). The two outgroup reporters (C. albicans (Ca) and S. 105 pombe (Sp)) transiently localized to the nucleus after the calcium induction and then continued 106 with stable nuclear localization during steady state ( Figure 2B). To quantify these phenotypic 107 differences at the single-cell level, we measured the duration and amplitude of reporter dynamics 108 by fitting Gaussian Processes to the single-cell trajectories (45). The results suggest that, 109 compared to the outgroup, Saccharomyces reporters have a shorter duration (represented by low 110 ln(l)) and higher amplitude (represented by high ln(a)) in their dynamics ( Figure 2C) Saccharomyces reporters quickly responded to calcium bursts on average ( Figure 2D). In 115 contrast, the average dynamics of the outgroup reporters are not affected by calcium bursts 116 ( Figure 2D).

117
Next, we estimated the amount of information encoded in the dynamics of pulsing 118 reporters. Information-theoretic approaches provide a natural framework to estimate the 119 information transmission capacity of cellular signaling pathways (47,48). We estimated the 120 mutual information between the dynamics of cytosolic calcium concentration and pulsing 121 reporters ( Figure 2E grey bars). We found that the dynamics of the Saccharomyces reporters 122 encoded more mutual information than that of the outgroup reporters (mean MI = 0.41 bits vs. 123 0.04 bits, 2-tails t-test, p =0.004, n = 9 and 6, respectively). By analyzing the inferred parameters 124 that describe the steady state dynamics, we also estimated mutual information between the 125 presence or absence of external calcium stress and the dynamics of pulsing reporters ( Figure 2E 126 white bars) and again found more mutual information in the Saccharomyces reporters (mean MI 127 = 0.72 bits vs. 0.13 bits, 2-tails t-test, p <10 -3 , n = 9 and 6, respectively). Taken together, these 128 results suggest that pulsing dynamics encode additional information about the environment and 129 that some functional sequence properties arose in the Crz1 IDR along the lineage leading to the 130 Saccharomyces.

131
Pulsing confers a fitness benefit in 0.2M calcium stress 132 If natural selection favored the evolution and preservation of pulsing along the lineage 133 leading to Saccharomyces, pulsing might confer a growth benefit. On the other hand, all the 134 orthologous IDRs support calcium-induced transient nuclear localization (Figure 2A and B, 135 (28,31)) and contain consensus calcineurin docking sites (49), a serine-rich NES region (28-31), 136 and several conserved phosphorylation sites (12) ( Figure 1C). Perhaps these are sufficient for 137 cell fitness, and pulsing is simply a non-functional elaboration of this phenotype. Therefore, we 138 sought to directly measure cell fitness under calcium stress (see Methods). To confirm that Crz1 139 function is needed for fitness in our assay conditions, we tested a mutant with CRZ1 deletion as 140 well as a mutant that conserved phosphorylation sites in the serine-rich NES region are mutated 141 ("mSRR" (38)). As expected, we found a large fitness defect for the CRZ1 deletion, and a small 142 but significant fitness defect for mSRR strain, confirming that our assay has the power to detect indicates that the benefit of pulsing is not simply due to high levels of nuclear localization during 147 steady states. 148 We next tested whether pulsing is beneficial to the cells by replacing the endogenous 149 IDRs of Crz1 with orthologous sequences. Consistent with the model where pulsing is preserved 150 because it is beneficial to the cell, we found that IDRs from the Saccharomyces clade, but not  Pulsing has been under stabilizing selection, but the underlying mechanisms have changed 161 We next asked if Crz1 pulsing can be found in the native systems of the outgroups. Based  Under that model, our finding above that the outgroup IDRs do not pulse in S. cerevisiae implies 169 that there must be compensatory changes that maintain pulsing in the outgroup species (42). To 170 distinguish between these models, we obtained strains ((29) and methods) of the two outgroup    189 Because only the IDRs of the Saccharomyces clade support pulsing, we asked which 190 parts of the IDRs are responsible for pulsing. Previous research showed that increasing the 191 affinity of one of the calcineurin docking sites, PxIxIT, leads to a higher pulsing frequency (25).

192
Therefore, we hypothesized that the PxIxIT strength of the Saccharomyces clade is higher than 193 its sister clade that includes C. albicans and that this increased strength is needed for pulsing. To   Figure 5A). This construct also contains two strong docking sites.

233
To determine if chimeric IDRs support pulsatility by responding to calcium bursts, we    The GFP-expressing S. pombe strain is a gift from Dr. Gordon Chua (29).

329
The C. albicans strain CaLC7415 with both copies of CRZ1 C-terminally tagged with NAT cassette, sgRNA, and Cas9 DNA were transformed into SN95. Upstream integration was 337 PCR tested using oLC600 and oLC9369, and downstream integration was tested using oLC274 338 and oLC9370. Lack of a wild-type allele was PCR tested using oLC9369 and oLC9373.  where means the number of generations and is the selection coefficient. 392 We found that this approach provides a resolution of the selection coefficient to 10 -3 and 393 successfully reproduced a previously reported small fitness defect (Supplementary Figure 3).

395
In the previous studies of pulsatile transcription factors, pulses were identified before 396 quantification and statistical analysis, e.g., pulse frequency(20,25) and pulse triggered averaging 397 (46). This approach presumes that the dynamics are pulsatile. However, in our case, whether a 398 pulsing reporter pulse or not was to be determined. Therefore, we needed a more general 399 approach to quantify single-cell trajectories. Shadi Zabad contributed several analyses that were ultimately not included in the manuscript. 466 We thank Drs P. Beltrao