Comprehensive Analysis of G1 Cyclin Docking Motif Sequences that Control CDK Regulatory Potency In Vivo

Abstract

Many protein-modifying enzymes recognize their substrates via docking motifs, but the range of functionally permissible motif sequences is often poorly defined. During eukaryotic cell division, cyclin-specific docking motifs help cyclin-dependent kinases (CDKs) phosphorylate different substrates at different stages, thus enforcing a temporally ordered series of events. In budding yeast, CDK substrates with Leu/Pro-rich (LP) docking motifs are recognized by Cln1/2 cyclins in late G1 phase, yet the key sequence features of these motifs were unknown. Here, we comprehensively analyze LP motif requirements in vivo by combining a competitive growth assay with deep mutational scanning. We quantified the effect of all single-residue replacements in five different LP motifs by using six distinct G1 cyclins from diverse fungi including medical and agricultural pathogens. The results uncover substantial tolerance for deviations from the consensus sequence, plus requirements at some positions that are contingent on the favorability of other motif residues. They also reveal the basis for variations in functional potency among wild-type motifs, and allow derivation of a quantitative matrix that predicts the strength of other candidate motif sequences. Finally, we find that variation in docking motif potency can advance or delay the time at which CDK substrate phosphorylation occurs, and thereby control the temporal ordering of cell cycle regulation. The overall results provide a general method for surveying viable docking motif sequences and quantifying their potency in vivo, and they reveal how variations in docking strength can tune the degree and timing of regulatory modifications.

Keywords: CDK; Cln2; SLiM; Sic1; Ste5; Whi5; cell cycle; cyclin; docking; phosphorylation.

Figure 1.. LP docking conservation among fungal G1 cyclins.

(A) Phylogeny of representative fungal genera and G1 cyclin types found in each subdivision. (B) Fungal G1 cyclins were expressed in S. cerevisiae as fusions to GST plus a half leucine zipper (GST-[lz]), along with HA-tagged substrates with no docking site, or an LP docking motif, or the partner half leucine zipper. Reduced mobility indicates phosphorylation. (C) Top, Cln1/2-CDK inhibits pheromone-induced MAPK activation and G1 arrest. Bottom, Ste20^Ste5PM chimeras contain a plasma membrane-binding domain (PM) and flanking CDK sites from Ste5 [11]; a functional LP docking motif allows these sites to be phosphorylated, which inhibits membrane binding by the PM domain and thus blocks signaling. Chimeras are tested in STE5–8A strains to prevent the CDK from inhibiting the native Ste5 protein [11, 22]. (D) P_GAL1-cyclin strains harboring Ste20^Ste5PM chimeras with either a functional LP motif (LP_Ste5) or a control sequence (nonLP) were treated with pheromone and tested for phosphorylation of MAPKs (Fus3 and Kss1). Below, quantified results (mean ± SEM, n = 4). (E) Phylogeny of fungal G1 cyclins [19] compared with LP docking abilities (check marks). Colored boxes denote three groups analyzed in panel F. (F)Surface residue conservation for three cyclin groups, overlaid onto a model for the Cln2-CDK complex (see Methods). The CDK is grey. Cyclin coloring reflects conservation: red, 100% identical; white, 90% identical; blue, 0% identical. The patch of strongest conservation includes residues previously implicated in LP docking [18]. Also see Figure S1.

Figure 2.. Extended LP motif sequences impart variation in cyclin recognition.

(A) LP motif sequences used throughout this study. (B) Four blocks of 4 residues in the Ste5 LP motif region were mutated (to AAAA) and tested for effects on substrate phosphorylation (as in Figure 1B) induced by Cln2 or Ccn1 cyclins. (C) Cln1/2 and Ccn1 cyclins were tested for phosphorylation of substrates with LP motifs from Ste5 or Sic1. Also see Figure S2. (D) P_GAL1-cyclin strains harboring Ste20^Ste5PM chimeras with different LP motifs were tested for pheromone-induced MAPK phosphorylation. Below, quantified results (mean ± SEM, n = 4).

Figure 3.. Competitive growth assay for the functional potency of cyclins and LP motifs.

(A) Schematic of the competitive growth assay, using Ste20^Ste5PM chimeras with a mixture of either wild-type (WT) LP motifs or libraries of randomized codon mutations. (B)Sequences of five LP motifs and the control (nonLP) motif; residues are colored where ≥ 50% are identical (blue) or similar (yellow). (C) Examples of competitive growth results showing changes in frequency of LP motifs after pheromone treatment, in cells ± P_GAL1-ScCLN2. Plots show averages of two independent experiments. Also see Figure S3A. (D) Experiments in panel C were performed in eight P_GAL1-cyclin strains and the control strain (none). Bars show the ratio (log₂) of the change in frequency for each motif compared to the nonLP motif (at 44 hr), normalized to the control strain; values show averages of two independent experiments. (E) Example showing changes in frequencies of Ste5 LP motif variants (in P_GAL1-ScCLN2 cells). The dashed black line is WT. Green and red show all WT synonyms and termination codons, respectively. Orange and blue show examples of missense synonyms with mild (Pro3 to Ser) or strong (Leu1 to Ser) defects; all six Ser codons are shown to illustrate that phenotypes are similar regardless of codon sequence. (F) Fitness score distributions for all nucleotide variants of two LP motifs (Ste5 and Sic1) in two P_GAL1-cyclin strains. Green, WT synonyms; Red, terminators; grey, missense mutations. Also see Figure S3B. (G) Correlation of fitness scores for all amino acid variants between two independent replicate experiments. Also see Figure S3C.

Figure 4.. Comprehensive analysis of LP motif sequence preferences.

(A) Fitness effects of all single-residue substitutions in five motifs. Red and blue indicate better and worse performance, respectively, relative to the wild-type motif; each panel is scaled to its own maximum and minimum. Diagonal lines depict standard errors, scaled such that the highest value covers the entire diagonal. Circles denote wild-type residues. Data represent three independent experiments. See Figure S4 for analyses with four other cyclins. (B) Sequence logos showing relative preferences of ScCLN2 along motifs in panel A. (C) Number of inactivating mutations (defined as normalized scores within 15% of minimum; see Methods) at each position in 4 “typical” motifs (i.e., excluding Whi5). Also see Figure S5. (D) Pearson correlation coefficients (r) for all pair-wise comparisons of cyclins, calculated from the raw fitness score matrices for all 5 motifs.

Figure 5.. Analysis of atypical and consensus motif preferences.

(A) Heat maps (as in Figure 4A) from mutational scanning of 11-residue sequences for four wild-type motifs plus two variants of the Whi5 motif (S-Whi5 and Whi5-P7L; magenta letters indicate sequence changes). Data represent two independent experiments. (B) Comparison of individual position preferences in five “typical” motifs. (C) Sequence logo showing the consensus preferences of the motifs in panel B.

Figure 6.. Observed and predicted potencies of new motifs.

(A) Frequency changes during competitive growth of 14 candidate LP motifs, plus Ste5 and nonLP controls, in cells with P_GAL1-ScCLN2 or no cyclin. Plots show averages of two independent experiments. (B) Panel A experiments were performed with eight P_GAL1-cyclin strains. Heat map colors indicate the ratio (log₂) of the change in frequency for each motif compared to the nonLP control (at 20 hr), normalized to the results in the no-cyclin strain; two independent experiments were averaged. Motif sequences are colored where >50% are identical (red) or similar (blue). (C) Observed vs. predicted potency of new motifs (excluding Ste5 and nonLP controls). Observed potency is the log₂ ratio value (as in panel B) for the ScCLN2 strain, at 20 or 44 hr. Predicted scores were derived from the preference score consensus PSSM (see Figures 5C and S6A) by calculating the sum of PSSM values for the residues at p1-p8 of each motif. Also see Data S1.

Figure 7.. A weak Sic1 motif delays its CDK-mediated degradation.

(A) Sic1 LP sequences from ten budding yeasts; residues are colored where ≥ 50% are identical or similar. Preferred residues are at top; arrows emphasize the conserved absence of favorites at p5 and p7. (B) Fluorescence intensities of Sic1-GFP variants, relative to the time when nuclear exit of Whi5-mCherry is 50% complete. Dark lines, mean; shaded bands, SEM. The “llpp” mutant substitutes VLLPP with AAAAA. (C) Degradation time and duration (see Methods) for Sic1-GFP variants. Circles, individual cells; lines, mean and 95% CI.