Budding marks initiation of cell division in Saccharomyces cerevisiae. Consequently, cell cycle progression can be monitored by the fraction of budded cells (budding index) in a proliferating cell population. We determined the budding index of a large collection of deletion strains, to systematically identify genes involved in cell cycle progression.
Saccharomyces cerevisiae divides asymmetrically. Once cells initiate cell division a small bud appears on the cell surface. This will eventually give rise to the daughter cell at the end of the cell cycle. Budding is an inherent property of S. cerevisiae cells that allows monitoring of cell cycle progression by simple morphological examination. In classic studies done by Hartwell and colleagues in the 70s (reviewed in ), this was used to identify conditional mutations in essential genes that gave rise to a uniform cell cycle arrest. Obviously, these early studies did not uncover any non-essential cell cycle regulators. Since bud emergence occurs soon after initiation of cell division, mutations in non-essential genes that alter the timing of cell cycle transitions may also alter the fraction of cells that are budded (budding index) in the culture. Here we present evidence that a systematic evaluation of the budding index of S. cerevisiae homozygous diploid deletion strains can be used to identify candidate genes involved in cell cycle progression.
2 Results and discussion
2.1 Rationale and outline of the analysis
Budding index represents the fraction of budded cells in an exponentially growing S. cerevisiae culture (i.e. Nbudded/Ntotal, where N is the number of cells). We evaluated the budding index of strains carrying deletions in non-essential genes, as a first criterion for identifying novel cell cycle regulators. The potential advantages of such an approach are simplicity and lack of bias. Instead of attempting to select for a rare mutant based on investigator-imposed criteria, in this case one examines mutant strains one at a time. The main disadvantage is that this process is time-consuming. Furthermore, although a cell cycle phenotype is usually accompanied by changes in the budding index, the converse is not always true, because budding index could be affected by numerous processes, not always directly related to cell cycle progression. To better interpret budding index measurements, one should also use the generation time of the culture. It should be pointed out that if all cell cycle transitions are equally affected in a generic manner, then a general slow growth phenotype may not be associated with a budding index phenotype. Furthermore, altering the timing of a cell cycle transition may not necessarily change the overall generation or doubling time (see Fig. 1). For example, deletion of gene products (e.g. G1 cyclins) that determine the timing of the G1/S transition results in a prolongation of the G1 phase. However, since there is a compensatory shortening of subsequent cell cycle phases there is no alteration in the overall growth rate or doubling time and, as a result, the culture displays a low budding index compared to wild-type cells. A comprehensive identification of deletion strains with a slow growth phenotype has been recently reported , facilitating the analysis of the budding index measurements we describe here.
Budding index in relationship to overall doubling time. Within one cell cycle, the G1 (unbudded) phase is indicated with solid blocks, and the S, G2, M (budded) phases are collectively indicated with open blocks. A lengthening of the G1 phase will not change the net doubling time if it is accompanied by a shortening of S, G2, M phases (i), but doubling time will increase if the S, G2, M phases remain unchanged compared to wild-type (ii). In (i), but not in (ii), a low budding index will be observed. Conversely, a shortening of the G1 (iii) will increase the budding index but it will not change the doubling time if it is accompanied by a compensatory lengthening of the S, G2, M phases. If, however, there is a block in S, G2, M (iv), there will be an increase in the budding index and the doubling time. Regarding G1 cyclin mutants, loss of function mutations resembles (i), while gain of function mutations resembles (iii).
We evaluated the budding index of a total of 1786 deletion strains. These strains were present in 19 96-well plates. The plates were selected at random from the homozygous diploid yeast deletion strain collection distributed by Research Genetics. The plate numbers that we have evaluated thus far, and for which we present the results below, are: 301–305, 308–311, 315–318, 323 and 330–334. For every eight strains that a budding index was obtained by any given investigator, the same investigator also measured the budding index of the wild-type strain, which was grown at the same time and under the same conditions as the deletion strains. Thus each investigator collected two groups of data, one for the deletion strains and one for the wild-type strain. The second group of data served as an internal control and was compared to the first group of that investigator. Deletion strains that had a budding index significantly different (by more than two standard deviations) from the wild-type group were classified as positive (Tables 1 and 2). Two different investigators examined each deletion strain, at different times, in the span of two years. Thus, we gave each deletion strain two chances to score as positive. Overall, 4.3% of the strains we examined displayed a budding index phenotype.
↵Strains that scored positive on both rounds of budding index evaluation are shown in bold.
2.2 Is the observed budding index of the deletion strains significantly different from the budding index of wild-type cells?
On Fig. 2A we have plotted the distribution of budding index measurements for wild-type and deletion strains. The variance of the wild-type plot represents the variance due to error and uncontrolled environmental factors. However, it is obvious that there is greater variance in the deletion strains. Therefore, although there is variability in budding index measurements, statistically significant differences are evident. These data also suggest that the trait in question, budding index, is complex and under multigenic control. Finally, budding index differences between deletion strains and wild-type are normally distributed (Fig. 2B), indicating a lack of bias in scoring a low vs. high budding index.
Deletion strains display greater variance in their budding indices than wild-type cells. The wild-type strain (BY4743, MATa/αhis3/his3 leu2/leu2 met15/met15 ura3/ura3) and the isogenic deletion strains were made by the Saccharomyces Genome Deletion Project, and obtained from Research Genetics/Invitrogen Co (Carlsbad, CA, USA). Budding index was evaluated during exponential growth in rich media (containing 1% yeast extract, 2% peptone, 2% dextrose) from at least 200 cells in each case. A: The budding index measurements were evaluated using Stata® release 6.0 software (College Station, TX, USA) to generate the summary statistics of the data. These parameters were introduced into the normal probability distribution function
where σ is standard deviation, x is the variable in question, and µ is the mean. Based on Eq. 1, we derived the curves for wild-type and deletion strains for each investigator, a representative of which is shown on A. The x-axis shows budding index, and the y-axis shows a standardized scale based on the standard deviation given by the formula
2Eq. 2 is a theoretical value applicable to all normal distributions, which increases as the standard deviation decreases, and vice versa. The curves were generated using Maple® release 5.0 software (Waterloo, Canada). B: Evaluation of the differences between the wild-type and the deletion strains. We generated the histogram shown on Fig. 1B, that encompasses the collective data from all investigators. On the x-axis the difference in budding index is shown, and the y-axis indicates the fraction of deletion strains with a corresponding budding index difference. The mean value for the differences was found to be µ=0.023 and σ=0.063. A 90% confidence interval on the mean differences was built and found to be 0.019, while the standard error of the mean for the confidence interval was 0.027. The differences were normally distributed, as tested by the Shapiro–Wilk test for normality, with W=0.98. This is shown by the overlaid graph of the normal probability distribution function, derived as in A.
2.3 Deletion strains with low budding index (Table 1)
The majority of the strains with low budding index carry deletions in genes encoding proteins involved in ribosome biogenesis or metabolism. This is not surprising given that initiation of cell division is very sensitive to the overall protein synthesis capacity of the cell . Overall, more than half (57%) of the deletion strains with a low budding index we identified grow poorly (see Table 1). We also identified strains deleted for genes of unknown function, which do not display a slow growth phenotype, and could be interesting candidates for future studies. Since only about 15% of all the homozygous diploid deletion strains displayed a slow growth phenotype , low budding index correlates with poor growth in general, consistent with the notion that growth requirements have to be met during the G1 phase of the cell cycle. Furthermore, it is worth pointing out that although about 10% of all deletion strains have significantly altered cell size , a disproportionate fraction (34%) of deletions with low budding index also had altered cell size (both larger and smaller than wild-type). All of these strains with low budding index and altered cell size also grow poorly (Table 1). It has been previously shown that slow growth due to defects in metabolism or protein synthesis results in small cell size . This is also evident in our data set, but we also find large cell size associated with slow growth and low budding index (Table 1). At least in some cases, this may be explained by the inability to generate a bud, such as the strains that lack BEM1 or BEM4 (Table 1).
2.4 Deletion strains with high budding index (Table 2)
In the strains listed on Table 2, the deleted open reading frames (ORFs) encoded gene products involved in a variety of biological processes, most of which when disrupted would be expected to result in an increase in budding index, suggesting again that our overall approach identifies physiologically relevant genes. These processes include bud growth, cell wall biogenesis, DNA synthesis and repair, mitosis and cytokinesis. Several of the genes listed on Table 2 have important and well-characterized roles in cell cycle progression, such as CLB2 that encodes a mitotic cyclin. Importantly, we also identified many genes of unknown function (Table 2), which could again be the target of future studies. Interestingly, compared to the strains with a low budding index, a significantly lower fraction of the strains with high budding index display a slow growth phenotype (56 vs. 19%, respectively), which is not significantly different from the fraction (15%) of all the homozygous diploid deletion strains with a slow growth phenotype . Likewise, the fraction of cell size alterations in deletion strains with high budding index was not significantly different from that of all deletion strains (12 vs. 10%).
2.5 Validation: VPS24 and VPS30
To further assess the validity of our approach we chose to examine two strains carrying deletions in genes involved in vacuolar (lysosomal) functions (VPS24 and VPS30) because at first they seemed very unlikely to have altered cell cycle progression, yet they scored as positive in our study. While VPS24 scored positive on both rounds of observation, VPS30 scored positive only once (Table 2). Both deletion strains do not proliferate slower than the wild-type strain (data not shown, and ). Based on DNA content analysis, deletion of either VPS24 or VPS30 increased the percentage of cells that have initiated DNA replication (Fig. 3A). Interestingly, the overall cell size of these strains was not reduced but it was instead slightly increased (Fig. 3B), explaining perhaps why these mutants had not been identified in previous cell size-based screens, where the ability to divide at a reduced size had been used to find genes that can accelerate initiation of DNA replication [4,5]. How Vps24p and Vps30p affect cell cycle progression is not clear. Interestingly, VPS30 is the yeast ortholog of human beclin 1, a tumor suppressor gene involved in autophagy . Beclin 1 functionally complements the autophagy defects of vps30Δ cells . We further confirmed the G1 shortening in synchronous vps30Δ cells obtained by elutriation (Fig. 3C). Cell cycle progression was measured based on the appearance of budded cells and the cellular DNA content, which was evaluated by flow cytometry (Fig. 3C). Note that the overall shortening of G1 was not modest, strongly supporting the validity of our overall approach. The molecular basis of beclin 1 tumor suppressor activity is unknown. It is also unknown whether cells lacking beclin 1 accelerate initiation of DNA replication, but based on our results this would seem a reasonable hypothesis.
A: Cell cycle profiles of vps24 and vps30 homozygous diploid strains. The cellular DNA content of the indicated strains was determined by flow cytometry . Cell numbers are plotted on the y-axis and the x-axis represents fluorescence. The percent of cells in G1 was quantified using Modfit® software, and the means from four independent experiments are shown. B: Cell size of live cultures of the indicated strains was measured using flow cytometry by forward angle scattering (FSC), shown on the x-axis, while cell numbers are plotted on the y-axis. The means from six independent experiments are shown. C: Early G1 cells of the indicated genotype were obtained by centrifugal elutriation, essentially as described previously . Briefly, cells grown in rich YPD media were loaded on a Beckman J6M/E centrifugal elutriator at a pump speed of 40 ml min−1 and elutriator rotor speed of 3200 rpm. The rotor speed was then lowered to 2400 rpm. Cells were collected at a pump speed of 52 ml min−1. The elutriated samples of the different strains shown had the same cell size at time 0. At the indicated time points after elutriation the cellular DNA content was determined by flow cytometry, and the percent of budded cells (%B) is also indicated. Cell numbers are plotted on the y-axis and the x-axis represents fluorescence.
In conclusion, we present a data set that will undoubtedly be a helpful starting point to investigators seeking to identify novel cell cycle regulators. It should be pointed out, however, that our analysis was based on the panel of deletion strains constructed by a large consortium, and even though we used diploid strains further confirmation of the gene disruption is needed before more detailed analysis of the cell cycle phenotype of a particular mutant. Overall, our approach was simple and relied exclusively on two easily obtained parameters: a morphological property (budding index) of S. cerevisiae, and generation time of liquid cultures. The availability of genomic information and panels of deletion strains for this organism made this study possible. Our results should give an accurate representation of the processes involved in determining budding index, since our analysis was extensive and covered 38% of the homozygous diploid deletion strains. Given the high degree of conservation between the yeast and human machineries that regulate cell division, information gathered from this study should be relevant to other organisms.
We thank J. Miller for flow cytometry and J. Witkowski, B. Parkinson, and A. Powers for help with budding index measurements. This work was supported by grants from the National Institutes of Health (GM062377) and the American Heart Association — Texas Affiliate (0060115Y) to M.P.
Michael F.Zettel, Lindsay R.Garza, Andrea M.Cass, Rachel A.Myhre, Lucy A.Haizlip, Shirley N.Osadebe, Daniel W.Sudimack, RituPathak, Thomas L.Stone, MichaelPolymenisFEMS Microbiol Lett(2003)223 (2):
253-258DOI: http://dx.doi.org/10.1016/S0378-1097(03)00384-7First published online: 1 June 2003 (6 pages)