Correspondence: Michael Polymenis, Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843, USA. Tel.: +1 979 458 3259; fax: +1 979 845 4946; e-mail: email@example.com
To identify Saccharomyces cerevisiae genes required for the proper timing of cell cycle transitions, we previously reported a systematic examination of the DNA content of homozygous diploid deletion strains. However, deletion strains with complex DNA content profiles were not examined in that study. Here, we report S. cerevisiae genes that when deleted give rise to DNA content profiles consistent with roles of the corresponding gene products during DNA replication. We also identified a set of genes whose deletion leads to increased DNA content, consistent with defects in mitosis, cytokinesis, or cell separation. Finally, we examined known interactions between the gene products of each group, placing these gene products in functional networks. Taken together, the data we present further validate the roles of the corresponding gene products in these processes, facilitating efforts to delineate gene function critical for genome replication, maintenance, and segregation.
DNA content measurements from asynchronous Saccharomyces cerevisiae cultures reflect the relative duration of cell cycle phases. DNA content analyses differentiate cells with unreplicated genome (in the G1 phase of the cell cycle) from cells replicating their DNA (in S phase), from cells with fully replicated genome (G2 or M phases), or from cells carrying extra amounts of DNA due to various defects in segregating the replicated genome to progeny.
In S. cerevisiae, DNA content analyses measured the effects of gene overexpression on cell cycle progression (Stevenson et al., 2001; Niu et al., 2008), cycle arrest when essential genes were turned-off (Yu et al., 2006), or effects on DNA replication in panels of deletion strains (Koren et al., 2010). Recently, we interrogated the yeast deletion collection of nonessential genes for altered DNA content, by flow cytometry (Hoose et al., 2012). Most strains displayed DNA content histograms with well-defined peaks, corresponding to cells with unreplicated (G1 phase of the cell cycle), or fully replicated genome (in the G2 or M phases of the cell cycle). Well-defined DNA content profiles allow for automated quantification of the percentage of cells in different phases of the cell cycle (Hoose et al., 2012). In that study, however, we did not analyze deletion strains with complex, unquantifiable DNA content profiles (Hoose et al., 2012). These DNA content profiles could arise from abnormal DNA replication, chromosome segregation, cytokinesis, or cell separation. Hence, information from DNA content analysis will be helpful to efforts aiming to understand these important cellular processes. Here, we present a survey for all the deletion strains that reproducibly displayed complex DNA content profiles. To our knowledge, such an analysis has not been reported previously.
Materials and methods
We queried the homozygous diploid deletion set (BY4743/MATa/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 LYS2/lys2Δ0 met15Δ0/MET15 ura3Δ0/ura3Δ0; obtained from Open Biosystems). All DNA content measurements were carried out from asynchronously dividing cells during exponential growth in rich media (YPD-1% yeast extract, 2% peptone, 2% dextrose), as we described elsewhere (Hoose et al., 2012). Detailed methods for flow cytometry and data acquisition have been published elsewhere (Hoose et al., 2012). All raw flow cytometry files were archived at Cytobank (Kotecha et al., 2010) and can be freely accessed, as we described previously (Hoose et al., 2012). The networks in Fig. 2 were constructed by summing cofunctionality scores from GeneMANIA composite networks (Warde-Farley et al., 2010) and displayed in Cytoscape using an edge-directed, force-generated layout (Smoot et al., 2011).
Elutriations and monitoring of synchronous cultures were performed as we have described elsewhere (Hoose et al., 2012). The ‘critical size’ is the size at which 50% of the cells have budded in these experiments (Hoose et al., 2012). To calculate ‘growth rate’ assuming exponential growth, we plotted the natural log (ln) of cell size as a function of time (in h). We fit the data to a straight line using the regression function in Microsoft Excel. From the slope of the line, we obtained the specific rate of cell size increase constant (k, in h−1). The average of all experiments for each strain was then calculated and shown in Fig. 3. Statistical comparisons of G1 variables were carried out with the unpaired Student's t-test function in Microsoft Excel, assuming unequal variance. Estimates of the length of G1 were calculated from G1(h) = ln(‘Critical Size’/‘Birth Size’)/k.
Results and discussion
We visually examined and manually curated each DNA content profile. We identified deletion strains that in at least two independent experiments reproducibly displayed complex DNA content profiles. In arranging the DNA content histograms, we took into account not only the overall appearance of each profile, but also the mean fluorescence intensity in each case (mean FL-A values, see Fig. 1).
Strains lacking URA7 (YBL039C), GLN3 (YER040W), DPB4 (YDR121W), RRM3 (YHR031C), CLB5 (YPR120C), or DIA2 (YOR080W) have a DNA content profile suggestive of abnormalities during DNA replication (Fig. 1a). Indeed, several of the corresponding gene products have well-established roles during DNA replication. Dpb4p is a subunit of DNA polymerase ε (Ohya et al., 2000). Rrm3p is a DNA helicase involved in DNA replication (Makovets et al., 2004). Clb5p is an S-phase cyclin (Bloom & Cross, 2007), while Dia2p is a protein that binds to origins of DNA replication (Koepp et al., 2006). Loss of the transcription factor Gln3p has been reported to lead to a DNA content profile consistent with cells accumulating in the S phase of the cell cycle (White et al., 2009), probably because Gln3p affects the expression of ribonucleotide reductase (Kwan et al., 2011). Because URA7 encodes the major subunit of CTP synthase involved in pyrimidine synthesis and maintenance of nucleotide pools (Ozier-Kalogeropoulos et al., 1991), loss of Ura7p may explain the accumulation of cells during the S phase of the cell cycle. Factors with related biological functions interact more often than expected by chance (Tong et al., 2004). We found that this is indeed the case for all the gene products in this set, including Ura7p (Fig. 2a), consistent with their DNA content profiles and putative roles in S phase.
An analogous systematic study using flow cytometry identified implicated all the above gene products in some aspect of DNA replication (Koren et al., 2010). The study by Koren et al. placed emphasis specifically on S phase, using higher-resolution DNA content analyses, possibly accounting for the fact that they identified twice as many (14 in total vs. 6 in our study) deletions with abnormal S-phase progression. In our analyses, we identified the deletions we mentioned above because their DNA content profiles were unquantifiable with the automated software we used (Hoose et al., 2012). Some deletions classified as S-phase mutants by Koren et al., resulted in well-defined DNA content profiles, leading us to classify them differently. For example, Koren et al. (2010) described S-phase defects in cells lacking the Cdk inhibitor Sic1p. While this is certainly true, the DNA content profile of sic1Δ cells is well defined and quantifiable, with a very prominent G2/M peak, enabling us to place sic1Δ cells in the group of mutants with a ‘low-G1’ DNA content (Hoose et al., 2012). Such differences in the analyses, together with the higher emphasis placed on S-phase progression by Koren et al., likely explain why we identified fewer gene products with roles in S phase.
In addition, Koren et al. (2010, fig. 1) described how a prolonged S phase was accompanied with a concomitant shortening of the preceding G1 phase. This was inferred from the DNA content profile, in conjunction with the growth rate measurements from asynchronous cultures. To examine this issue in more detail, we examined synchronous cultures obtained by centrifugal elutriation. The amount of time cells spend in G1 depends on several variables. As we have described elsewhere (Hoose et al., 2012), the absolute length of the G1 phase can be measured if one knows three parameters: (1) The size of newborn cells (‘birth’ size); (2) the ‘critical size’ these newborn daughter cells must attain to initiate cell division; and (3) the rate (‘growth rate’) at which they progress from their birth size to their critical size (Fig. 3a). We obtained each of these variables for cells lacking Dpb4p and Ura7p. We found that cells lacking Dpb4p have no defects in growth rate, and they divide at a size similar to that of wild-type cells (Fig. 3). However, because dpb4Δ cells are born significantly larger than wild-type cells do (49.8 fl vs. 40.6 fl, P =6.1 × 10−11, based on a Student's t-test), the G1 phase of daughter dpb4Δ cells is shorter (0.8 h, vs. 1.4 h for wild-type cells). This is in agreement with Koren et al., and consistent with the specific and direct role of Dpb4p in DNA replication, with no growth defects in dpb4Δ cells. Next, we examined cells lacking Ura7p. In this case, we found that ura7Δ cells have a 50% longer G1 phase than the G1 phase of wild-type cells. Specifically, ura7Δ cells grow slower (k =0.23 h−1 vs. 0.28 h−1 for wild-type cells, P =0.007, based on a Student's t-test). They also have to reach a larger critical size (69.5 fl vs. 61.5 fl for wild-type cells, P =0.0005, based on a Student's t-test) before they initiate a new round of cell division (Fig. 3). Hence, in contrast to the interpretation of Koren et al., inferred from asynchronous cultures, we demonstrate that the absolute length of the G1 phase is substantially increased in cells lacking Ura7p. This is consistent with the broader roles of Ura7p in the maintenance of nucleotide pools, expected to affect a multitude of cellular processes in addition to DNA replication. We summarize schematically our results in Fig. 3e. G1 estimates of different mutants have to take into account the fact that different mutants may start and finish at different sizes and progress in G1 at different rates. The methods of Koren et al. did not evaluate in detail such parameters. Hence, a diminished G1 peak in DNA content profiles is not necessarily accompanied by an absolute shortening of the G1 phase.
Next, we identified a set of deletions strains in which a fraction of cells has DNA content higher than a fully replicated diploid genome (Fig. 1b). The genes deleted in these strains include several with known roles in cytokinesis or cell separation. Elm1p (YKL048C) is a protein kinase involved in cytokinesis (Bouquin et al., 2000). Ace2p (YLR131C) is a transcription factor important for the destruction of the septum after cytokinesis (Sbia et al., 2008). Finally, while Def1p (YKL054C) is a chromatin-associated protein with roles in transcription elongation, Def1p has also been implicated in cytokinesis (Jordan et al., 2007). Other gene products in this group have already been reported to function in chromosome maintenance and segregation. Ypk1p (YKL126W) is a pleiotropic protein kinase with roles during progression through the G2 phase of the cell cycle (Sopko et al., 2006). Est1p (YLR233C) is a telomere homeostasis factor (Zhang et al., 2010). Cnm67p (YNL225C) is a spindle pole component (Schaerer et al., 2001). Spc72p (YAL047C) is a component of γ-tubulin that binds spindle pole bodies (Hoepfner et al., 2002). Ctf4p (YPR135W) is required for sister chromatid cohesion (Hanna et al., 2001). Mms22p (YLR320W) is a subunit of an E3 ubiquitin ligase with roles in DNA replication and repair and chromosome segregation (Hanna et al., 2001). Finally, Apn1p (YKL114C) is a DNA repair endonuclease (Boiteux & Guillet, 2004). Although the evidence linking these gene products with the above processes often did not include a DNA content analysis, there is also a multitude of interactions among the genes in this group (Fig. 2b), consistent with their shared roles.
In summary, the DNA content profiles of the deletion strains we examined here provide additional phenotypic support for the function(s) of the corresponding gene products in genome replication, maintenance, segregation, and cell separation. Similar comprehensive studies may reveal analogous gene products involved in these processes in other organisms, including humans.
The authors have no conflict of interest to declare. This work was supported by a grant from the National Science Foundation (MCB-0818248) to M.P.