The Snf3p glucose sensor is required for inhibition of sporulation by glucose
To measure glucose inhibition of meiosis, we transferred growing cells to nutrient medium containing acetate, which promotes meiosis, and 0%, 0.5%, or 2% glucose (Sp, Sp+0.5, and Sp+2 medium, respectively). After 72 h, wild-type yeasts sporulated fivefold less efficiently in Sp+0.5 than in Sp, and 50-fold less efficiently in Sp+2 ().
Fig. 1 Involvement of Rgt2p and Snf3p glucose sensors in repressing sporulation in the presence of glucose. Wild-type (WT, SH1232), rgt2Δ (SH2402), snf3Δ (SH2431), rgt2Δsnf3Δ (SH1926), mth1Δ (SH3993), rgt2Δsnf3Δmth1Δ (more ...)
We next tested the role of the Rgt2p and Snf3p glucose sensors in regulating meiosis. For this purpose, we transferred rgt2Δ, snf3Δ, and rgt2Δsnf3Δ strains to Sp, Sp+0.5, and Sp+2 media and determined the proportion of these cultures that formed asci after 72 h (). As expected, all three mutants sporulated to approximately the same level as the wild type in Sp medium. In addition, in Sp+0.5 medium the rgt2Δ mutant sporulated at a frequency comparable to the wild type, consistent with the fact that the Rgt2p sensor only responds to high glucose concentrations. In contrast, the snf3Δ mutant (and the rgt2Δsnf3Δ double mutant) sporulated approximately three times more efficiently than the wild type in Sp+0.5; indeed, these two mutants sporulated nearly as efficiently in Sp+0.5 as in Sp. Thus Snf3p, which senses moderate concentrations of glucose, is required for this concentration of glucose to inhibit sporulation.
At a higher concentration of glucose (Sp+2), spore formation was strongly repressed in all of the above strains. Consistent with the fact that both Rgt2p and Snf3p are activated by high glucose, the rgt2Δsnf3Δ double mutant sporulated to significantly higher levels than the wild type (unpaired t-test, P = 0.046). The snf3Δ mutant may also sporulate to higher levels than the wild type, although the difference was not quite significant (P = 0.06). Because these experiments do not distinguish whether both Rgt2p and Snf3p are required for high glucose to repress spore formation or only Snf3p is required for this repression, most of our subsequent experiments were performed in the rgt2Δsnf3Δ double mutant. Importantly, even in this double mutant, sporulation in Sp+2 is less than one-sixth as efficient as in Sp. Thus, pathways that do not require Rgt2p/Snf3p sensors must also mediate inhibition of meiosis at high glucose concentrations.
Snf3p inhibits sporulation during late stages of growth in colonies
The Sp+0.5 medium may mimic conditions encountered during late stages of growth on glucose, when both a low concentration of glucose and a higher concentration of nonfermentable carbon source are present. To measure the role of glucose sensors in repressing meiosis during late stages of growth in colonies, we used a genetic assay that detects meiotic cells on the surface of colonies (Purnapatre & Honigberg, 2002
). In brief, colonies grown on the medium containing glucose were transferred together with the underlying nutrient agar medium to a medium containing the drug canavanine. Canavanine, in the strain background used for these experiments, selects specifically for meiotic cells. As a result, meiotic cells on the surface of the original colony give rise to smaller satellite colonies after transfer to the new medium.
Wild-type colonies incubated on SC medium lacking arginine (2% glucose) for 8 days did not yield satellite colonies after transfer and neither did rgt2Δ colonies; in contrast, snf3Δ and snf3Δrgt2Δ colonies grown under the same conditions efficiently form satellite colonies (). As expected, when snf3Δrgt2Δ and wild-type colonies were suspended in water and examined by light microscopy, at least several-fold more asci were observed in the mutant than the wild type (2% vs. < 0.5%, respectively). Thus, the glucose sensors are required to efficiently repress sporulation in colonies grown on SC medium.
Fig. 2 Involvement of Rgt2p and Snf3p in regulating meiosis in colonies grown on glucose medium. Colonies of the wild-type (WT), rgt2Δ, snf3Δ, and rgt2Δsnf3Δ strains used in were grown for 8 days on SC medium and then transferred (more ...)
We next examined the role of the glucose sensors in regulating sporulation in colonies grown on one of two other carbon sources: acetate and raffinose. Because acetate is a nonfermentable carbon source, the glucose sensors would not be expected to affect growth (Bisson et al., 1987
) or sporulation on this medium. Indeed, wild-type and rgt2Δsnf3Δ
colonies grew at approximately the same rate on the 2% acetate medium, and these two strains sporulated to almost the same frequency (after 8 days, wild-type colonies yielded 11.6 ± 4.5% asci, whereas rgt2Δsnf3Δ
colonies yielded 11.1 ± 2.8% asci). In contrast to acetate, raffinose is a fermentable carbon source. Efficient metabolism of raffinose requires the Rgt2p/Snf3p glucose induction pathay, at least in part, because this pathway mediates the induction of invertase (SUC2
), which converts raffinose to glucose (Ozcan et al., 1997
). Consistent with earlier results, we found that rgt2Δsnf3Δ
colonies grew more slowly than wild-type colonies on raffinose medium (Neigeborn & Carlson, 1984
). In addition, rgt2Δsnf3Δ
colonies sporulated at much higher levels than wild-type colonies (after 8 days, wild-type colonies yielded 0.1 ± 0.1% asci, whereas rgt2Δsnf3Δ
colonies yielded 14.1 ± 1.2% asci). Thus, the glucose sensors repressed meiosis in colonies on carbon sources (glucose and raffinose) that activate the glucose induction pathway but not on a carbon source (acetate) that does not activate this pathway.
Glucose sensors not required to repress IME1 and IME2 transcription
Because glucose inhibits transcription of IME1 and IME2, we asked whether the glucose sensors are required for this repression. Wild-type and rgt2Δsnf3Δ strains were grown to midlog phase, transferred to Sp, Sp+0.5, or Sp+2 medium, and then assayed for IME1 and IME2 transcripts at various times (). As expected from the experiments shown in , in Sp medium the rgt2Δsnf3Δ double mutant expressed both the IME1 and the IME2 transcript to the same levels as the wild type (, compare lanes 1–4 to lanes 11–14). Also consistent with , in Sp+2 medium both the wild type and the rgt2Δsnf3Δ mutant expressed IME1 only at the same basal level as in stationary-phase cells and did not express detectable IME2 even after 20 h incubation (, lanes 1, 8–11, and 18–20). Because the rgt2Δsnf3Δ mutant sporulates at detectable (although low) levels after 72 h in Sp+2 (), presumably, IME1 and IME2 are eventually induced in at least some cells in the culture.
Fig. 3 Role of Rgt2p and Snf3p in regulating IME1 and IME2 transcript levels. Wild-type (WT, SH1232) and rgt2Δsnf3Δ (SH1926) cultures were grown in YPA and transferred to sporulation media as in . At the indicated times after transfer (0, (more ...)
As described above, there was no effect of deleting the glucose sensors on IME1 and IME2 transcript levels in either Sp or Sp+2 media; however, there is a slight effect of the deletions on IME2 induction in Sp+0.5. In the wild type, IME1 induction was slightly delayed and IME2 induction clearly delayed in Sp+0.5 relative to Sp; however, eventually the IME2 transcript reached approximately the same high levels in both media (, compare lanes 2–7). The rgt2Δsnf3Δ mutant induced IME2 transcript in Sp+0.5 earlier than in wild type, although not as early as in Sp (, compare lanes 12–14 to lanes 15–17). Thus in Sp+0.5, the glucose sensors have only a modest effect on the timing of IME2 transcript accumulation.
In summary, RGT2 and SNF3 are not required for high glucose to repress IME1 and IME2 and are only partially required for moderate glucose to delay IME2 induction. Thus, although SNF3 efficiently inhibits spore formation when moderate concentrations of glucose are present, it is unlikely that this sensor inhibits spore formation by repressing either IME1 or IME2 transcription.
Glucose sensors regulate Ime2p turnover
As described in the Introduction, SCFGrr1p
is required both for the glucose induction pathway and for the glucose-stimulated turnover of Ime2p (Purnapatre et al., 2005
). To determine whether the glucose sensors also regulate Ime2p turnover, we compared Ime2p-6XHA stability in wild-type and rgt2Δsnf3Δ
, driven by the tetO promoter, was induced in the growth medium and then repressed by the addition of tetracycline as cells were transferred to Sp or Sp+0.5 medium. As expected from our previous study (Purnapatre et al., 2005
), in wild-type cells, Ime2p-6XHA was stable in the Sp medium but rapidly turned over in the Sp+0.5 medium (, compare lanes 2–4 with lanes 5–7). In contrast, in the rgt2Δsnf3Δ
mutant, Ime2p-6XHA was stable in both the Sp and Sp+0.5 media. Thus, the glucose sensors are required for glucose-stimulated Ime2p turnover, and it is likely that glucose sensors inhibit sporulation, at least in part, by triggering Ime2p degradation.
Fig. 4 Effect of glucose sensors and the Rgt1p transcription factor on Ime2p-6HA stability. Wild-type (WT, SH3354), rgt2Δsnf3Δ (SH3343), and rgt1Δrgt2Δsnf3Δ (SH3342) strains were grown in SC-Leu to midlog, and then transferred (more ...)
Snf3p mediates glucose inhibition of spore formation by negatively regulating Mth1p
Once bound to glucose, the glucose sensors inactivate two transcriptional repressors, Std1p and Mth1p. Of these two repressors, Mth1p is most important in maintaining repression of the GIGs, such as hexose transporters and other genes involved in responding to glucose (Schmidt et al., 1999
; Kim et al., 2006
). Thus, when glucose is present in the media, Mth1p is inactivated and genes responding to glucose are induced. To determine whether glucose sensors inhibit sporulation by inactivating Mth1p, we measured the efficiency of spore formation in mth1Δ
mutants (). Because Mth1p represses the response to glucose, we did not expect the single mth1Δ
mutant to affect sporulation in Sp medium, and as expected, sporulation levels in this mutant were not significantly different from the wild type (P
= 0.15). Similarly, sporulation in Sp+0.5 or Sp+2 media was not very different in the mth1Δ
mutant compared with the wild type, consistent with Mth1p being repressed by glucose. Furthermore, the mth1Δrgt2Δsnf3Δ
triple mutant sporulated with nearly the same efficiency as the rgt2Δsnf3Δ
double mutant in Sp medium. However, the triple mutant sporulated much less efficiently than the double mutant in Sp+0.5 or Sp+2 medium. Indeed, in this latter mutant, the triple mutant was inhibited at least as efficiently as the wild type in either medium. This result indicates that at moderate concentrations of glucose, Snf3p likely inhibits meiosis by inactivating Mth1p.
Because the glucose sensors are known to target Mth1p for destruction by activating yeast casein kinase 1, we attempted to determine whether this kinase was also required for glucose to inhibit meiosis. Yeast casein kinase 1 is an essential protein encoded by two redundant genes YCK1
(Robinson et al., 1993
). A yck1Δyck2ts
strain is viable at 25 °C but not at 37 °C. Unfortunately, in the strain background used, even the wild type was defective in sporulation at temperatures > 33 °C, a temperature that is still semi-permissive for growth in the mutant (not shown). Thus we were unable to determine whether yeast casein kinase 1 is involved in glucose repression of sporulation.
The rgt1Δ mutant is defective in meiosis in the absence of glucose, and this defect correlates with small cell size
Mth1p induces transcription of HXTs
and other genes by triggering the dissociation of the Rgt1p transcriptional repressor from the promoters of these genes (see Introduction). To determine whether the glucose sensors repress sporulation through this same pathway, we initially examined sporulation in the rgt1Δ
mutant. Surprisingly, this mutant sporulated only half as efficiently as the wild type in Sp medium (i.e. in the absence of glucose), an unexpected result given that Rgt1p is only known to regulate glucose-responsive genes (). However, a possible explanation for this result was suggested by the observation that rgt1Δ
cells were on an average 30% smaller than wild-type cells (). Indeed, a number of mutants that display small cell size also exhibit defects in meiosis (Calvert & Dawes, 1984
; Zhang et al., 2002
; Day et al., 2004
), and cell size may be a principal determinant of the timing of meiosis under optimal conditions (Nachman et al., 2007
). To test the idea that the rgt1Δ
mutant was defective in meiosis because of its size, an rgt1Δ
mutant and a wild-type strain were grown as above, separated by centrifugal elutriation, and cells with different sizes isolated and transferred to a sporulation medium (). As with wild-type cells, larger rgt1Δ
cells sporulated much more efficiently than smaller cells; indeed large rgt1Δ
cells sporulated to levels comparable to the wild type. Thus, the sporulation defect of the rgt1Δ
mutant is likely attributable to its smaller size.
Fig. 5 Sporulation defect in the rgt1Δ mutant correlates with cell size. (a) Wild-type and rgt1Δ cultures were grown as in and sizes in the cell population determined. (b) Cultures were grown as in and cell fractions of the indicated (more ...)
Snf3p regulates spore formation but not Ime2p turnover through Rgt1p
Because the rgt1Δ mutant has only a partial defect in spore formation, we could still measure the effect of glucose on sporulation in this mutant. As expected, in Sp+2 medium, the rgt1Δ mutant, like the wild type, failed to form spores (<2%) even after 72 h of incubation (). However, the rgt1Δ mutant was somewhat more sensitive to moderate glucose concentrations than was the wild type [(sporulation in Sp)/(sporulation in Sp+0.5) equals 5 for the wild type and 8.8 for the mutant]. This result is consistent with genes required for glucose signaling or transport being released from repression in the rgt1Δ mutant.
As a more definitive test of the role of Rgt1p in regulating meiosis, we measured spore formation in an rgt1Δrgt2Δsnf3Δ triple mutant. Because sporulation in the rgt2Δsnf3Δ mutant was less sensitive to glucose than wild type, whereas sporulation in the rgt1Δ mutant was more sensitive, the rgt1Δrgt2Δsnf3Δ triple mutant was used to determine the epistasis relationship between the rgt1Δ and rgt2Δsnf3Δ mutants. In Sp medium, the triple mutant, like the rgt1Δ single mutant, sporulated approximately twofold less efficiently than the wild type (). More importantly, the triple mutant, again like the rgt1Δ single mutant, was more sensitive to 0.5% glucose than the wild type [(sporulation in Sp)/(sporulation in Sp+0.5) equals 13.3]. Thus, rgt1Δ is epistatic to rgt2Δsnf3Δ with respect to glucose inhibition of sporulation, indicating that the glucose sensors inhibit spore formation at 72 h at least in part by repressing Rgt1p activity.
To determine whether the glucose sensors stimulate Ime2p degradation by repressing Rgt1p, we compared Ime2p turnover in the rgt2Δsnf3Δ double mutant with an rgt1Δrgt2Δsnf3Δ triple mutant (). If glucose-stimulated turnover of Ime2p requires inhibition of Rgt1p, then the triple mutant should restore glucose-stimulated turnover. Surprisingly, we found no difference in Ime2p turnover between the double and triple mutants. Thus, the sensors do not stimulate Ime2p turnover solely by inhibiting Rgt1p function. A corollary of this conclusion is that inhibition of spore formation by moderate glucose, because it does involve Rgt1p, must not depend on destabilizing Ime2p.