Specific Disruption of Ras2 CAAX Proteolysis Alters Its Localization and Function

Abstract

Many CAAX proteins, such as Ras GTPase, undergo a series of posttranslational modifications at their carboxyl terminus (i.e., cysteine prenylation, endoproteolysis of AAX, and carboxylmethylation). Some CAAX proteins, however, undergo prenylation-only modification, such as Saccharomyces cerevisiae Hsp40 Ydj1. We previously observed that altering the CAAX motif of Ydj1 from prenylation-only to canonical resulted in altered Ydj1 function and localization. Here, we investigated the effects of a reciprocal change that altered the well-characterized canonical CAAX motif of S. cerevisiae Ras2 to prenylation-only. We observed that the type of CAAX motif impacted Ras2 protein levels, localization, and function. Moreover, we observed that using a prenylation-only sequence to stage hyperactive Ras2-G19V as a farnesylated and nonproteolyzed intermediate resulted in a different phenotype relative to staging by a genetic RCE1 deletion strategy that simultaneously affected many CAAX proteins. These findings suggested that a prenylation-only CAAX motif is useful for probing the specific impact of CAAX proteolysis on Ras2 under conditions where other CAAX proteins are normally modified. We propose that our strategy could be easily applied to a wide range of CAAX proteins for examining the specific impact of CAAX proteolysis on their functions. IMPORTANCE CAAX proteins are subject to multiple posttranslational modifications: cysteine prenylation, CAAX proteolysis, and carboxylmethylation. For investigations of CAAX proteolysis, this study took the novel approach of using a proteolysis-resistant CAAX sequence to stage Saccharomyces cerevisiae Ras2 GTPase in a farnesylated and nonproteolyzed state. Our approach specifically limited the effects of disrupting CAAX proteolysis to Ras2. This represented an improvement over previous methods where CAAX proteolysis was inhibited by gene knockout, small interfering RNA knockdown, or biochemical inhibition of the Rce1 CAAX protease, which can lead to pleiotropic and unclear attribution of effects due to the action of Rce1 on multiple CAAX proteins. Our approach yielded results that demonstrated specific impacts of CAAX proteolysis on the function, localization, and other properties of Ras2, highlighting the utility of this approach for investigating the impact of CAAX proteolysis in other protein contexts.

Keywords: CAAX pathway; CAAX proteolysis; Ras; Ras signaling; Ras2; Rce1; lipidation; palmitoylation; proteolysis; shunt pathway; yeast.

FIG 1

Posttranslational modifications of CAAX proteins. (A) Initial isoprenylation of a consensus cysteine amino acid within the carboxyl-terminal CAAX motif yields a farnesylated (C₁₅; shown) or geranylgeranylated (C₂₀) protein. Shunted CAAX proteins are not further modified. Canonical CAAX proteins are subject to additional endoproteolysis and carboxylmethylation. Some canonical CAAX proteins, including certain Ras isoforms, are also modified with a palmitoyl lipid on a nearby cysteine (e.g., yeast Ras2, human NRas, and human HRas). CAAX proteolysis is not required for palmitoylation of mammalian HRas, and this issue was investigated for Ras2 in this study. (B) Ras2-CAAX variants evaluated in this study and their expected PTMs.

FIG 2

Impact of CAAX motifs on Ras2 viability function. The 5-FOA plasmid-loss assay was performed with yeast initially coexpressing Ras1 from URA3-marked plasmid and the indicated Ras2-CAAX variant from a LEU2-marked plasmid. Growth on 5-FOA medium, indicative of the functional ability of the Ras2-CAAX variant, was assessed by qualitative patching (upper) or quantitative colony counting (bottom). For the qualitative test, cells were first patched on YPD solid media, on which cell growth could be supported by either Ras1 or a functional Ras2-CAAX variant, and then replica plated onto 5-FOA solid medium, on which cell growth could only be supported by a functional Ras2-CAAX variant. For quantitative assessment, strains cultured in SC-Ura, Leu were plated at low density on YPD solid medium to determine CFU, and resultant colonies were replica plated onto 5-FOA solid medium to determine the number of cells that could be supported by the indicated Ras2-CAAX variant. Percent growth was determined by deriving the ratio of colony counts on the 5-FOA to YPD plates for each strain and comparing the ratio to that of Ras2-CIIS, which was set to 100%. Values for the data shown in the graph were averaged from at least 4 independent experiments having both biological and technical replicates, and the error bars depict standard errors of the means. Statistical analysis was performed using one-way analysis of variance and Bonferroni’s post hoc test. *, P < 0.001; NS, not significant. The yeast strain used was RJY510; plasmids used were pRS315, B250, Ras2-Scaax, Ras2-SsaaX, pWS1613, and pWS1615.

FIG 3

Impacts of CAAX motifs on Ras2 farnesylation and palmitoylation. (A) Yeast cells expressing the indicated Ras2-CAAX variants were cultured in SC-Ura, Leu, and whole-cell lysates prepared by alkaline lysis were analyzed by SDS-PAGE (6% stacking; 10% resolving) and anti-Ras2 immunoblotting. A duplicate image of the immunoblot (lower panel) was modified with lines centered on each band for better visualization of mobility differences. Ponceau S staining was used to visualize loading across lanes (see Fig. S1 in the supplemental material). The strain used was RJY510; plasmids used were pRS315, B250, Ras2-Ssaax, Ras2-Scaax, pWS1613, and pWS1615. (B) The acyl-PEG exchange assay was performed on yeast cell lysates containing Ras2-CAAX variants. Equivalent amounts of each lysate were treated with (+) or without (−) hydroxylamine (HAM) followed by 5-kDa mPEG-Mal prior to evaluation by SDS-PAGE and immunoblotting with anti-Ras2 antibody. The mPEG-Mal-labeled and unlabeled populations are indicated (closed and open triangles, respectively). The values for percent labeled were derived by taking the ratio of labeled to total signal in each lane. The labeling pattern was reproducible across at least 2 independent experiments; a representative result is shown. The strain used was RJY510 transformed with plasmids B250, Ras2-Scaax, and pWS1613; transformants were selected on 5-FOA to express only the Ras2 isoform.

FIG 4

Impacts of CAAX motifs on Ras2 protein levels and GTP loading. (A) Yeast cells grown in SC-Ura, Leu liquid medium for setting up the 5-FOA plasmid-loss assay (see Fig. 2) were used in parallel for whole-cell lysate preparation by alkaline lysis. Short- and long-time exposures of the anti-Ras2 immunoblot are shown. Equivalent loading was assessed by Ponceau S staining (see Fig. S2A in the supplemental material). The strain used was RJY510; plasmids used were pRS315, B250, Ras2-Scaax, Ras2-Ssaax, pWS1613, and pWS1615. For quantification, immunoblot band intensities were evaluated using ImageJ from at least 4 independent experiments, inclusive of panel A data, that reflected different cell growth and extract preparation conditions (see Fig. S2B to D). (B) Late-log-phase cultures of yeast cells grown in SC-Leu liquid medium were treated with 250 μg/mL cycloheximide (t = 0 min), and samples were collected at the indicated time points for whole-cell lysate preparation by alkaline lysis. Samples were analyzed by anti-Ras2 and anti-Kar2 immunoblotting. The strain used was RJY510; plasmids used were B250, Ras2-Scaax, Ras2-Ssaax, pWS1613, pWS1615, and pWS1890; transformants were selected on 5-FOA to only express the Ras2 isoform, except for the transformant expressing Ras2-SSIIS, which only grows when Ras1 is coexpressed. (C) The Ras pulldown activation assay was performed using a commercial kit on cell lysates prepared by bead-beating of cells. Equivalent amounts of each lysate were used for the assessment of input Ras2 (total Ras2) and activated Ras2 (Ras2-GTP), which were confirmed by Ponceau S staining (see Fig. S2E). Ratios of Ras2-GTP to total Ras2 relative to Ras2-CIIS were calculated for multiple exposures across 2 experiments, except for SSIIS, whose ratio was calculated for multiple exposures from a single experiment. Average ratios are indicated below each lane; error bars depict standard deviations. The strains used were as described for panel B.

FIG 5

Impact of CAAX motifs on Ras2 localization. (A) Total cell lysates of strains expressing Ras2-CAAX variants were prepared by bead-beating and used for differential fractionation. Equivalent percentages of each sample representing the total lysate (T), supernatant (S), and particulate (P) fractions from 100,000 × g centrifugation were evaluated by immunoblotting with anti-Ras2 antibody. The percentage of signal associated with the P fraction relative to the sum of signal from P and S fractions is noted, and values were obtained from at least 2 independent experiments. The strain used was RJY510; plasmids used were B250, Ras2-Scaax, Ras2-Ssaax, pWS1613, and pWS1615. (B) Fluorescence microscopy was performed with the GFP-tagged Ras2-CAAX variants under fluorescence optics (63×). The image shown for each variant is representative of at least 30 fields of cells. The strain used was LRB938; plasmids used were pWS1714, pWS1735, pWS1821, pWS1823, and pWS1889.

FIG 6

Impact of CAAX motifs on Ras2-G19V activity. (A and B) The acute heat shock assay was performed using yeast expressing the indicated Ras2-CAAX variants. Saturated cultures grown in SC-Leu medium were normalized for density, heated (55°C, 10 min) or mock treated (no heat shock), and serially diluted 1:5 before pinning on SC-Leu (heat-shocked) and YPD (no heat shock) plates. Yeast strains used were LRB938 and yWS3126; plasmids used were pRS315, B250, B561, B562, pWS1612, pWS1614, and pWS1890. (C) The cdc25(TS) temperature sensitivity assay was performed using yeast transformed with Ras2-CAAX variants. The strains were cultured in SC-Leu liquid medium, normalized for density, serially diluted 1:10, and pinned on two SC-Leu plates that were incubated at permissive room temperature (RT) and restrictive temperature (37°C) up to 5 days. The yeast strain used was cdc25-2; plasmids used were pRS315, B250, B561, B562, pWS1612, pWS1614, and pWS1890. For panels A to C, the data shown are representative of the patterns of growth observed across at least 3 independent experiments having both biological and technical replicates; representative results are shown. (D) Whole-cell lysates were prepared by alkaline lysis and analyzed as described for Fig. 4A. Equivalent portions of each lysate were evaluated, except for Ras2-G19V-CIIS, which was evaluated in a higher amount to facilitate immunoblot detection (see Fig. S1). Short- and long time-exposures of the anti-Ras2 immunoblot are shown. The strain used was RJY510; plasmids used were B561, B562, pWS1612, pWS1614, and pWS1890.

FIG 7

Model relating PTMs to properties of Ras2. (A) Qualitative representations of phenotypes for viability function (see Fig. 2), protein levels (total and GTP-activated) (see Fig. 4A and C, respectively), and Ras2 localization (see Fig. 5) observed for Ras2-CAAX variants examined in this study. The scaling is relative to the Ras2-CAAX variant, with the weakest phenotype observed within the set for each assay. (B) Our experimental observations were used to model the observed effects of PTMs. The model proposes that all Ras2-CAAX variants are GTP-activated but have different functional abilities due to varying effector engagements. Fully modified Ras2-CIIS (i.e., farnesylated, cleaved, and palmitoylated) is at the plasma membrane, has the lowest protein levels, and has highest signaling intensity; the widths of solid arrow(s) emanating from the membrane-bound Ras2/effector complex are indicative of intensity; the length and direction of dashed arrows represent the likelihood of a preferred state of membrane association or effector engagement. Partially modified Ras2-SCIIS and Ras2-CASQ (i.e., farnesylated and either CAAX cleaved or palmitoylated, respectively) are mislocalized from PM, have increased protein levels, and have reduced signaling intensity. Farnesylated-only Ras2-SCASQ and unmodified Ras2-SSIIS are mislocalized into the cytosol, have the highest protein levels, and have the lowest signaling intensities. We propose that increased Ras2 levels (total and GTP-activated) are a compensatory cellular response to increased overall Ras2 signaling output when loss of PTMs leads to reduced Ras2 signaling; the number of solid arrows is representative of this effect but not proportional to actual ratios.