Genome-Wide Screen for Haploinsufficient Cell Size Genes in the Opportunistic Yeast Candida albicans

One of the most critical but still poorly understood aspects of eukaryotic cell proliferation is the basis for commitment to cell division in late G1 phase, called Start in yeast and the Restriction Point in metazoans. In all species, a critical cell size threshold coordinates cell growth with cell division and thereby establishes a homeostatic cell size. While a comprehensive survey of cell size genetic determinism has been performed in the saprophytic yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe, very little is known in pathogenic fungi. As a number of critical Start regulators are haploinsufficient for cell size, we applied a quantitative analysis of the size phenome, using elutriation-barcode sequencing methodology, to 5639 barcoded heterozygous deletion strains of the opportunistic yeast Candida albicans. Our screen identified conserved known regulators and biological processes required to maintain size homeostasis in the opportunistic yeast C. albicans. We also identified novel C. albicans-specific size genes and provided a conceptual framework for future mechanistic studies. Interestingly, some of the size genes identified were required for fungal pathogenicity suggesting that cell size homeostasis may be elemental to C. albicans fitness or virulence inside the host.

haploinsufficiency to identify genes and biological process that influence size control in C. albicans. Given the importance of C. albicans as an emerging eukaryotic model, very little is known regarding the genetic networks that control size homeostasis in this opportunistic yeast. A systematic screen using elutriation-based size fractioning (Cook et al. 2008) coupled to barcode sequencing (Bar-seq) identified 685 genes (10% of the genome) that influenced size control under optimal growth conditions. While C. albicans and S. cerevisiae share the morphological trait of budding, and core cell cycle and growth regulatory mechanisms (Berman 2006;Cote et al. 2009), a limited overlap was obtained when comparing the size phenome of both yeasts. This genome-wide survey will serve as a primary entry point into the global cellular network that couples cell growth and division in C. albicans.

Combination of C. albicans mutants into a single pool
A sterilized 384-well pin tool was used to transfer DBC mutant cells into Nunc Omni Trays containing YPD-agar, and colonies were grown for 48 hr at 30°. Missing or slow growing colonies were grown separately by repinning 3.5 ml from the initial liquid cultures. Each plate was overlaid with 5 ml of YPD, and cells were resuspended using Lazy-L spreader and harvested by centrifugation for 5 min at 1800 · g. The obtained cell pellet was resuspended in 20 ml fresh YPD and DMSO was added to 7% (v/v). Mutant pools were aliquoted and stored at 280°.

Cell size selection by centrifugal elutriation
The mutant pool was size-fractioned using centrifugal elutriation with the Beckman JE-5.0 elutriation system. This technique separates cells on the basis of size. A tube of pooled mutant population was thawed on ice and used to inoculate 2 l of YPD at an OD 595 of 0.05. Mutant cells were grown for four generations at 30°under agitation to reach 5 · 10 10 cells. Cells were then pelleted by centrifugation and resuspended in 50 ml fresh YPD. To disrupt potential cell clumps and separate weakly attached mother and daughter cells, the 50 ml pooled cells were gently sonicated twice for 30 sec. The resuspended cells were directly loaded into the elutriator chamber of the Beckman JE-5.0 elutriation rotor. A 1 ml sample of cells was retained separately as a pre-elutriated cell fraction. The flow rate of the pump was set to 8 ml/min to ensure the loading of cells. To elute small cell size mutant fractions, the pump flow rate was increased in a step-wise fashion (in 2-4 ml/min increments). For each flow rate, a volume of 250 ml was collected from the output line of the rotor.

Bar-seq
Bar-seq was performed using Illumina HiSeq2500 platform. Genomic DNA was extracted from each cell fraction using YeaStar kit (Zymo Research). The 20-bp UpTag barcode of each strain were amplified by PCR (Xu et al. 2007). Primers used for PCR recognize the common region of each barcode and contain the multiplexing tag and sequences required for hybridization to the Illumina flow cell. PCR products were purified from an agarose gel using the QIAquick Gel Extraction kit (Qiagen) and quantified by QuantiFluor dsDNA System (Promega). Bar-seq data were processed as following: after filtering out low frequency barcode counts, the complete set of replicate barcode reads were normalized using a cyclic loess algorithm (R package "limma"). Reads from individual elutriation fractions, relative to the pre-elutriation population, were further M-A loess normalized and converted to Z scores.

Confirmation of cell size phenotypes
Cell size determination was performed using a Z2-Coulter Counter channelizer (Beckman Coulter). The Coulter principal is based on electrical impedance measurement, which is proportional to cell volume (Coulter 1953). C. albicans cells were grown overnight in YPD at 30°, diluted 1000-fold into fresh YPD and grown for 5 hr at 30°to reach a final density of 5 · 10 6 -10 7 cells/ml, a range in which size distributions of the different WT strain used in this study do not change. A total of 100 ml of exponentially growing cells was diluted in 10 ml of Isoton II electrolyte solution, sonicated three times for 10 sec and the distribution measured at least three times on a Z2-Coulter Counter. Size distribution data were normalized to cell counts in each of 256 size bins and size reported as the peak median value for the distribution. Data analysis and size distribution visualization were performed using the Z2-Coulter Counter AccuComp software.

Determination of critical cell size
Critical sizes of cln3/CLN3, cdc28/CDC28 and sch9/SCH9 mutants were determined using budding index as a function of size. G1 daughter cells were obtained using the JE-5.0 centrifugal elutriation system (Beckman Coulter) as described previously (Tyers et al. 1993). C. albicans G1-cells were released in fresh YPD medium and fractions were harvested at an interval of 10 min to monitor bud index. Additional fractions were collected to assess transcript levels of the RNR1 and ACT1 as cells progressed along the G1 phase.

Real-time quantitative PCR
A total of 10 8 G1 phase cells were harvested, released into fresh YPD medium,, grown for 10 min prior to harvesting by centrifugation and stored at 280°. Total RNA was extracted using the RNAeasy purification kit (Qiagen) and glass bead lysis in a Biospec Mini 24 bead-beater, as previously described (Sellam et al. 2009). cDNA was synthesized from 2 mg of total RNA using the SuperScript III Reverse Transcription system [50 mm Tris-HCl, 75 mm KCl, 10 mm dithiothreitol, 3 mm MgCl 2 , 400 nm oligo(dT) 15 , 1 m random octamers, 0.5 mm dNTPs, and 200 U Superscript III reverse transcriptase]. The total volume was adjusted to 20 ml, and the mixture was then incubated for 60 min at 42°. Aliquots of the resulting first-strand cDNA were used for real-time quantitative PCR (qPCR) amplification experiments. qPCR was performed using the iQ 96-well PCR system for 40 amplification cycles and QuantiTect SYBR Green PCR master mix (Qiagen). Transcript levels of RNR1 were estimated using the comparative Ct method, as described by Guillemette et al. (2004), and the C. albicans ACT1 open reading frame as a reference. The primer sequences were as follows: RNR1-forward: 59-GACTATCTACCATGCT GCTGTTG-39; RNR1-reverse: 59-GGTGCAACCAACAAGGAGTT-39; ACT1-forward: 59-GAAGCCCAATCC AAAAGA-39; and ACT1-reverse: 59-CTTCTGGAGCAACTCTCAATTC-39.

Gene ontology analysis
Gene ontology (GO) term enrichment of size mutants was determined using the Generic GO Term Finder tool (http://go.princeton.edu/cgi-bin/ GOTermFinder), with multiple hypothesis correction (Boyle et al. 2004). Descriptions related to gene function in Supplemental Material, Table S2 were extracted from the Candida Genome Database (CGD) database (Inglis et al. 2012). Information related to gene essentiality/dispensability was taken from O' Meara et al. (2015) and the CGD database.

Data availability
The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article.

RESULTS AND DISCUSSION
The Cln3-Cdc28 kinase complex and Sch9 are haploinsufficient for cell size In S. cerevisiae, a number of critical Start regulators are haploinsufficient for cell size, including the rate-limiting G1 cyclin Cln3 and a number of essential ribosome biogenesis factors, such as the AGC kinase Sch9 (Sudbery et al. 1980;Jorgensen et al. 2002). To test whether size haploinsufficiency exists in C. albicans homologs, size distributions of the heterozygous mutants of AGC kinase Sch9, the cyclin Cln3 G1, and its associated cyclin-dependent kinase Cdc28 were examined. Both cln3/CLN3 and cdc28/CDC28 showed an increase of size as compared to their congenic parental strain, with median sizes 13% (59 fl) and 19% (62 femtolitre) larger than the WT strain (52 fl), respectively ( Figure  1A). As in S. cerevisiae, sch9/SCH9 exhibited a reduced size of 23% (40 fl) as compared to WT.
Two hallmarks of Start, namely SBF-dependent transcription and bud emergence, were delayed in both cln3/CLN3 and cdc28/CDC28 and accelerated in sch9/SCH9, demonstrating that the Cln3-Cdc28 complex and Sch9 regulate the cell size threshold at Start. The cln3/CLN3 mutant passed Start after growing to 92 fl, 24% higher than the parental WT cells, which budded at 74 fl ( Figure 1B). Similarly, cdc28/CDC28 reached Start at 105 fl, which is 41% higher than WT. The onset of G1/S transcription was delayed in both mutants, as judged by the expression peak of the G1-transcript RNR1 ( Figure 1C). The small mutant sch9/SCH9 passed Start at 30 fl, a size 60% smaller than the WT, and displayed accelerated G1/S transcription (Figure 1, B and C). These data demonstrate that, as in S. cerevisiae, size haploinsufficiency in C. albicans can be used to screen for dosage-dependent regulators of growth and division at Start.
A high-throughput screen for cell size haploinsufficiency To identify all dosage-sensitive regulators of size in C. albicans, a genomewide screen was performed where pooled mutants were separated based on their size by centrifugal elutriation and their abundance determined by Bar-seq. This method has been previously validated in S. cerevisiae (Cook et al. 2008), yielding a high degree of overlap when compared to a strain-by-strain analyses (Jorgensen et al. 2002) (Figure 2A). In the current study, we screened a comprehensive set of 5470 heterozygous deletion diploid strains from the Merck DBC collection (Xu et al. 2007) for cell size defects. This collection covers 90% of the 6046 proteincoding open reading frames based on the current CGD annotation (Binkley et al. 2014). Two small cell size fractions were obtained by centrifugal elutriation and were used for these experiments ( Figure  2B). Small cells and corresponding small deletion mutants are enriched in these fractions, while large cells strains are depleted. To determine mutant abundance in each fraction, genomic DNA of each pool was extracted and barcodes were PCR-amplified and sequenced. Abundance of each mutant in each fraction was appreciated by calculating the ratio of elutriated cells counts over counts of pre-elutriated cells.
To identify mutants with size defects, a two-step filter was applied. First, a size cut-off value was determined based on a benchmark set of conserved small (sch9/SCH9) and large (cln3/CLN3 and cdc28/CDC28) size mutants for which size was reduced or increased at least 12% as compared to the parental WT strain. Second, a normalized Z-score of 1.5 and 21.5 was used to identify both small (whi) and large (lge) size mutants, respectively. A total of 12 size mutants were excluded from our analysis, since they were found in both whi and lge datasets (Table S1). Microscopic examination revealed that these mutants had a remarkable size heterogeneity and grew predominantly as pseudohyphae. Based on these criteria, we identified 685 mutants that exhibited a size defect in both elutriated fractions. This includes 382 whi and 303 lge mutants (Table S2). As expected, cln3/CLN3 and cdc28/CDC28 mutants were identified as lge mutants, while sch9/SCH9 was found among the smallest mutant in the elutriated pools. A total of 15 whi and 15 lge mutants were randomly selected and their size was measured by electrolyte displacement on a Coulter Z2 channelizer. The obtained data confirmed size defect in all 30 mutants examined (Table S3). Size phenotype of two heterozygous mutants, including the rac1/RAC1 (Bassilana and Arkowitz 2006) and sec15/SEC15 (Guo et al. 2016) previously shown as lge mutants, were confirmed by our analysis. However, the protein kinase C pkc1/PKC1 mutant exhibited a whi phenotype in our investigation, while it was previously identified as large size (Paravicini et al. 1996). To clear up this contradiction, we have created new pkc1/PKC1 mutants. At least five independent transformants were sized and the whi phenotype was confirmed for all of them (data not shown).
Lge mutants were predominantly defective in functions related to the mitotic cell cycle ( Figure 2D and Table S2). These mutants include genes required for G1/S transition (G1 cyclin Cln3 and Ccn1, Cdc28 and Met30) suggesting that delay in G1 phase is the primary cause of their increased size. We also found that mutations in processes related to DNA replication (ORC3, ORC4, MCM3, CDC54, RFC3, PIF1, SMC4, ELG1), G2/M transitions (Hsl1, Cdc34) and cytoskeleton-dependent cytokinesis (MYO5, INN1, SEC15, CDC5, CHS1) conferred an increase of cell size. A similar observation was reported in S. cerevisiae, where a recent genomewide microscopic quantitative size survey uncovered that mutants of the G2/M transition and mitotic exit fail to properly control their size. The large size of cell cycle mutants support the fact that cell growth and cell cycle are separate processes and cells continue to grow and increase their size without commitment to divide. Other investigations propose a model where, in addition to the G1-phase, size is sensed and controlled at G2/M checkpoint (Anastasia et al. 2012;King et al. 2013;Harvey and Kellogg 2003;Soifer and Barkai 2014). However, further analysis will be necessary to provide further insights into the presumptive linkage of each phase of the cell cycle and size homeostasis in C. albicans.
While a large proportion of whi mutants in C. albicans were related to ribosome biogenesis, inactivation of genes controlling translation initiation (ASC1, SCD6, PAB1, GCD6, GCD2, SUI1, EIF4E, GCD11) resulted in lge phenotype. A similar finding was reported in different genome-scale surveys of size phenome in S. cerevisiae (Jorgensen et al. 2002;Soifer and Barkai 2014). This large size phenotype in these mutants could be explained by the fact that regulators of Start onset, such as G1 cyclin Cln3 (Barbet et al. 1996;Polymenis and Schmidt 1997), are sensitive to the rate of translation initiation.
Plasticity of size phenome and C. albicans fitness Recent evidence has uncovered an extensive degree of rewiring of both cis-transcriptional regulatory circuits and signaling pathways across many cellular and metabolic processes between the two budding yeasts, C. albicans and S. cerevisiae (Lavoie et al. 2010;Li and Johnson 2010;Blankenship et al. 2010;Lavoie et al. 2009;Sellam and Whiteway 2016). In S. cerevisiae, a similar size haploinsufficiency screen was performed in heterozygous diploid strains of essential genes (Jorgensen et al. 2002). To assess the extent of conservation and plasticity of the size phenome, genes that were haploinsufficient for cell size in C. albicans were compared to their corresponding orthologs in S. cerevisiae. This analysis revealed a limited overlap between the two species with five whi (rpl18a, sch9, rlp24, nop2, nog1) and two lge (rpt4, cln3) mutants in common. In fact, genes with reciprocal size phenotypes were similar in frequency (the whi mutants rpt2/RPT2 and pkc1\PKC1 in C. albicans had lge phenotype in S. cerevisiae).
Interestingly, the corresponding homozygous deletion mutants of many C. albicans haploinsufficient size genes were shown to be re-quired for virulence. A total of 69 size genes (representing 10%), including 47 small and 22 large size mutants, in our dataset were linked to C. albicans virulence or adaptation in the human host ( Figure 2E). This suggests that cell size is an important virulence trait that can be targeted by antifungal therapy. Hypothetically, virulence defect in small size mutant could be linked to the reduced surface of the contact interface between C. albicans, with either host cells or medical devices in case of biofilm infections. Indeed, we have previously shown that the whi transcription factor mutant ahr1 had attenuated virulence and exhibited a decreased attachment ability to abiotic surface such polystyrene, which consequently impaired biofilm formation (Askew et al. 2011). On the other hand, virulence defect in lge mutant could be associated with the fact that cells with large surfaces had a decreased lifespan which might impact their fitness and their viability inside the host (Yang et al. 2011;Mortimer and Johnston 1959).
While the link between C. albicans size and virulence remains uncharacterized, many investigations reported that many other fungal pathogens such as Cryptococcus neoformans and Mucor circinelloides adjust their cell size to access to specific niche in the host or to escape from immune cells (Wang and Lin 2012). In C. albicans, recent investigations have shown that large gastrointestinally induced transition cells, as compared to the standard yeast form, define the commensal form of this fungus (Pande et al. 2013). Furthermore, Tao et al. (2014) recently uncovered a novel intermediate phase between the White and C. albicans mating competent opaque phenotypes, called the Gray phenotype. The Gray cells are similar to opaque cells in general shape, however, they exhibit a small size and low mating efficiency. The Gray cell type has unique virulence characteristics, with a high ability to cause cutaneous infections and a reduced capacity in colonizing internal organs such as kidney, lung, and brain. Taken together, these lines of evidence emphasize the possible link between cell size and C. albicans fitness. Figure 1 The Cln3-Cdc28 kinase complex and the AGC kinase Sch9 control Start in C. albicans. (A) Size distributions of the WT strain (CAI4) as compared to lge mutants cln3/CLN3 and cdc28/CDC28, as well as the whi mutant sch9/SCH9. (B and C) Start is delayed in cln3/CLN3 and cdc28/ CDC28 and accelerated in sch9/SCH9. (B) Elutriated G1 phase daughter cells were released into fresh media and monitored for bud emergence as a function of size. (C) G1/S transcription. RNR1 transcript level was assessed by quantitative real-time PCR and normalized to ACT1 levels.
In summary, we provided the first comprehensive genome-wide survey of haploinsufficient cell size in a eukaryotic organism. In contrast to S. cerevisiae, where a similar screen was limited to essential genes (Jorgensen et al. 2002), our screen spanned the genome. A total of 300 (43.8%) dispensable genes and only 87 (12.7%) essential genes were haploinsufficient for size. Overall, our screen identified known conserved regulators (Sch9, Sfp1, Cln3) and biological processes (ribosome biogenesis and cell cycle control) required to maintain size homeostasis in this opportunistic yeast. We also identified novel C. albicans sizespecific genes and provided a conceptual framework for future mechanistic studies. Interestingly, some of the size genes identified were required for fungal pathogenicity, suggesting that cell size homeostasis may be elemental to C. albicans fitness or virulence inside the host.  Size distributions of pooled DBC mutants before (pre-elutriated) and after elutriation of two small cell fractions (28 and 34 ml/min). (C and D) GO terms enrichment of whi (C) and lge (D) mutants (P . 1e205). GO analysis was performed using GOTermFinder (http:// go.princeton.edu/cgi-bin/GOTermFinder). (E) Overlap between C. albicans genes haploinsufficient for cell size and those affecting virulence phenotypes. Avirulent mutant phenotypes were obtained from CGD based on decreased competitive fitness in mice and/or reduced invasion and damage to host cells.