Supplemental Data

Full Text
Systematic Genetic Analysis with Ordered Arrays of Yeast Deletion Mutants
Amy Hin Yan Tong, Marie Evangelista, Ainslie B. Parsons, Hong Xu, Gary D. Bader, Nicholas Pagé, Mark Robinson, Sasan Raghibizadeh, Christopher W. V. Hogue, Howard Bussey, Brenda Andrews, Mike Tyers, and Charles Boone

Supplementary Material

View Supplemental Table 2

Materials and Methods
SGA Analysis: Manual and Robotic Arraying Procedur
For arraying by hand, yeast cells were transferred with a 96 Floating-Pin Replicator (VP408FH), containing 96 pins of 1.58mm-diameter, and a colony Copier (VP380), which allows for construction of a 3-by-3 array of 96 colonies (V & P Scientific, Inc.; The strains were arrayed in triplicate, resulting in 3X 96 grid of 288 different strains and a total of 864 colonies/plate. To sterilize the replicator, it was first placed in a tray of water (~2 min), which removes most of the yeast cells from the pins. Next, the replicator was placed in a tray of 10% bleach (20 sec), followed by 3 sequential rinses in different water baths (10 sec/bath). Finally, the replicator was placed in 95% ethanol and then flamed. For automated arraying, yeast cells were transferred using a 768 floating-pin replicator, a design based upon the methodology developed for automated two-hybrid screens developed in Stan Fields' laboratory []. To sterilize the pins, first they were placed in a tray of water (2 min); second, the pins were placed in a sonicator containing 70% ethanol (2 min); third, they were placed in a tray of 95% ethanol (30 sec); finally, the pins were fan-dried (30 sec). Automated arraying was carried out using a Virtek Colony Arrayer ( 1mm-diameter pins were used in all the steps except for the final transfer, for which 0.5mm-diameter pins were used.

SGA Analysis: Media and Growth Conditions
The deletion mutant array (DMA) was propagated on rich medium containing G418 (200mg/L;InVitrogen). The following procedure facilitated SGA analysis. First, query strains were grown overnight in 5ml of rich medium (YEPD), then transferred to an empty Omnitray. To generate a source of newly grown query cells for mating to the DMA, the query strain culture was pinned onto rich (YEPD) plates (at a density of 384/plate for hand-pinning and 768/plate for automated pinning) and the cells were grown for one day. Second, for the mating reaction, the query strain was first pinned onto a fresh YEPD plate, the DMA was pinned on top of the query cells, and the plates were incubated at room temperature for one day. The resulting MATa/name zygotes were pinned onto medium that selects for growth of diploid cells; for example, if the query strain carried a natR-marked mutation, then diploids were selected on rich medium (YEPD) containing G418 (200 mg/L) and clonNAT (100mg/L;Werner BioAgents). If the query strain carried a URA3-marked mutation, then diploids were selected on synthetic dextrose medium (SD) lacking uracil and lysine. Third, diploid cells were pinned to sporulation medium (2% agar, 1% potassium acetate, 0.1% yeast extracts, 0.05% glucose, supplemented with uracil, histidine, and leucine) and the arrays were incubated for five days at 22°C. Fourth, to select for growth of MATa spore progeny, spores were pinned on to haploid selection medium [SD medium lacking histidine and arginine but containing canavanine :SD - His/Arg + canavanine (50mg/l)] and incubated at 30�C (2 days). Fifth, the MATa cells were repinned to haploid-selection medium for a second round of selection and incubated at 30�C for 1 day. Sixth, to select for meiotic progeny that carry the deletion mutation derived from the DMA parental strain, MATa haploids were pinned onto SD medium containing G418 and grown for 1 day. Because ammonium sulfate impedes the function of G418 and clonNAT, synthetic medium containing these antibiotics was made with monosodium glutamic acid as a nitrogen source [20g/L agar, 1.7g/L yeast nitrogen base w/o ammonium sulphate and amino acids (Difco), 1g/L monosodium glutamic acid, 2g/L amino acid drop-out lacking histidine and arginine, 50mg/L canavanine, and 200mg/L G418]. Finally, prior to scoring colony size, double mutants were selected on synthetic haploid-selection medium containing G418 and clonNAT (100mg/L clonNAT) for two days. For temperature sensitive mutants, the mating and haploid selection steps were carried out at a permissive temperature (~26°C).

SGA Analysis: Yeast Strains and Reporters
Derivatives of BY4741 (MATaura3name0 leu2name0 his3name1 met15name0) and BY4742 (MATnameura3name0 leu2name0 his3name1 lys2name0) were used to construct strains designed specifically for this study [Brachmann et al., Yeast14, 115 (1998)]. Two different starting strains were used for the SGA screens, Y2454 (MATnamemfa1name::MFA1pr-HIS3can1name ura3name0 leu2name0 his3name1 lys2name0) and Y3068 (MATnamecan1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0). In Y2454, HIS3 was integrated at the MFA1 locus such that it is regulated by the MFA1 promoter (pr), mfa1name::MFA1pr-HIS3. In Y3068, MFA1pr-HIS3 was integrated at the CAN1 locus, can1name::MFA1pr-HIS3. The construction of Y2454 involved three steps. First, to create Y2402 (MATnamecan1nameura3name0 leu2name0 his3name1 lys2name0) which carries can1name, BY4742 was transformed with Bam HI-linearized p3630 and transformants were selected on SD-URA media; Ura+ colonies were streaked on 5-FOA plates to select for loss of the integrated URA3 maker and the pRS306 backbone; integration of can1name was confirmed by canavanine resistance. To create p3630, CAN1 was amplified from yeast genomic DNA with primers (5'-GCGAACAGAGTAAACCGAA-3' and 5'-GAAGGTCTGAAGGAGTTC-3'), cloned into pRS306, then digested with HindIII and religated, to remove a 1345-bp fragment within the CAN1 gene. Second, to create Y2420 (MATa mfa1name::MFA1pr-HIS3 ura3name0 leu2name0 his3name1 met15name0), the MFA1 open reading frame was replaced by the HIS3 gene via integration of a PCR product generated with primers (5'-ATCTGTAACTGTTTCTCGGATAAAACCAAAATAAGTACAAAGCCATCGAATAGAAATGACAGAGCAGAAAGCCCT-3') and (5'-GGTGTAGCGGAAAAGGAAGATAAAGGAGGGAGAACAACGTTTTTGTACGCAGAAACTACATAAGAACACCTTTGG-3'), which anneal to HIS3 and contain MFA1 sequences (in bold), and pEG202 plasmid DNA [J. Gyuris, E. Golemis, H. Chertkov, R. Brent, Cell75, 791 (1993)] as a template. Third, to create Y2454 (MATnamemfa1name::MFA1pr-HIS3can1name ura3name0 leu2name0 his3name1 lys2name0), Y2420 was crossed to Y2402, the resulting MATa/name diploids were sporulated, and spore progeny with the appropriate markers were recovered. The construction of Y3068 involved four steps. First, a 455-bp fragment containing the upstream sequence of CAN1 was amplified from yeast genomic DNA with primers (5'-TAGGGCGAACTTGAAGAATAACC-3') and (5'-GCCACGTTGCACACTATCCTGTGCTATGCCTTTTTTTTTTTTTGTT-3'), which contained a 21-bp sequence (in bold) from the 5' end of MFA1pr-HIS3. Second, MFA1pr-HIS3 was amplified from genomic DNA obtained from Y2420 with primers (5'CAGGATAGTGTGCAACGTGGC3') and (5'-ATCAAAGGTAATAAAACGTCATATCTACATAAGAACACCTTTGGTGG-3'), which contained a 24-bp sequence (in bold) from the 5' end of the CAN1 downstream sequence. Third, a 300-bp fragment containing the CAN1 downstream sequence was amplified from yeast genomic DNA with primers (5'-ATATGACGTTTTATTACCTTTGAT-3') and (5'-ACGAAAAATGAGTAAAAATTATCTT-3'). Finally, the set of PCR products above were used as templates to generate a fused product with primers (5'-TAGGGCGAACTTGAAGAATAACC-3') and (5'-ACGAAAAATGAGTAAAAATTATCTT-3'). BY4741 transformants carrying can1name::MFA1pr-HIS3 were selected on synthetic medium lacking histidine. Y2927 (MATname bni1name::natRmfa1name::MFA1pr-HIS3can1name ura3name0 leu2name0 his3name1 lys2name0) was created in two steps. First, Y2454 was first transformed with a BamHI-NotI cut p1041, which carries a bni1name::URA3 allele deleted for an internal ~3.6 kb MscI-HpaI BNI1 fragment, to generate Y2613 (MATname bni1name::URA3mfa1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0). Second, bni1name::URA3 was switched to bni1name::natR by PCR-based integration with primers (5'-AGTATTCTTAACCCAACTGCACAGAACAAAAACCTGCAGGAAACGAAGATAAATCATGACCACTCTTGACGACACGG-3' and 5'-TTGAAGCTCTAATTTGTGAGTTTAGTATACATGCATTTACTTATAATACAGTTTTCTAGGGGCAGGGCATGCTCAT-3'), which anneal to natMX4 DNA [A. L. Goldstein, J. H. McCusker, Yeast15, 1541 (1999)] and contain URA3 sequences (in bold). Y3122 (MATname bbc1name::natRmfa1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0) was created in two steps. First, BBC1 was deleted from Y2454 using PCR-based integration with primers (5'-CGCGTCCCAACTGCAAGAGCTACTAGAAGTAAACACAGTCTTACCCAAACTAAGGAGGAAACAGCTATGACCATG-3' and 5'-GGAAAATGATGAAATCGAAAGCATTCACAATTTCCCCTTGGAAGCATTGGGTAAAGTTGTAAAACGACGGCCAGT-3'), which annealed to URA3 and contained BBC1 sequences (in bold), to generate Y2980 (MATnamebbc1name::URA3mfa1name::MFA1pr-HIS3can1nameura3name0 leu2name0 his3name1 lys2name0). Second, bbc1name::URA3 was switched to bbc1name::natR as described above. The construction of Y3029 (MATnamearc40-40::URA3mfa1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0) involved four steps. First, PCR-mutagenized DNA [R. C. Cadwell, G. F. Joyce, PCR Methods Appl3, S136 (1994)] carrying ARC40 and 200bp downstream of its stop codon was amplified with primers, (5'- ATGTCATTTTCCAATTCTAAAGACAA-3') and (5'- GGGACGAGGCAAGCTAAACAGATCTTCTTTGGGTAGCAATTCTAAATAAT-3'), which contains 25bp (in bold) complementary sequence to the URA3MX6 cassette. Second, URA3MX6 was amplified from p4348 using primers (5'-AGATCTGTTTAGCTTGCCTCGT-3') and (5'- AGGTAGGGAATACACATAACAATATATATATATACGTATGTATATGAATTCGAGCTCGTTTAAACTGGA-3'), which contains 45bp (in bold) complementary to the sequences downstream of ARC40. Third, Y2454 was co-transformed with the two PCR products described above. Fourth, transformants were selected on SD medium lacking uracil and replica-plated to 37°C to identify temperature-sensitive alleles of ARC40. To create p4348 (pCRII-TOPO::URA3MX6), the URA3MX6 cassette was amplified from pJJ224 (Rosetta Inpharmatics) with primers (5'-AGATCTGTTTAGCTTGCCTCGT-3' and 5'-GAATTCGAGCTCGTTTAAACTGGA-3') and ligated into pCRII-TOPO (InVitrogen). The construction of Y3181 (MATname arp2-33::URA3mfa1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0) involved five steps. First, ARP2 (with 200bp downstream sequence) was amplified using mutagenic PCR (R. C. Cadwell, G. F. Joyce ibid pS136) from yeast genomic DNA using primers, which contains 25bp sequence (in bold) that anneals to the URA3MX6 cassette (5'-ATGGACCCACATAATCCAATTGG-3' and 5'-GGGACGAGGCAAGCTAAACAGATCTACTGAATACTTGAAATACCAAGAAAA-3'). Second, URA3MX6 was amplified from p4348 using primers, which contains 45bp sequence (in bold) that anneals to the 45bp downstream sequence of ARP2 (5'-AGATCTGTTTAGCTTGCCTCGT-3' and 5'- ATTTAAACTTTTTTATTAATTCTTTTGGATCAATAAAGTTCAAATGAATTCGAGCTCGTTTAAACTGGA-3'). Third, Y2454 was transformed with the two PCR products described above. Fourth, Ura+ transformants were selected on SD-URA and finally screened for temperature sensitivity at 37°C. We used a switcher/replica-plating method (see below) to create the following strains Y3304 (MATnamebim1name::natR mfname1name::MFname1pr-LEU2 can1name::MFA1pr-HIS3 ura3name0 leu2name0 his3name1 lys2name0), Y3310 (MATnamesgs1name::natRmfname1name::MFname1pr-LEU2 can1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0) and Y3334 (MATnamerad27name::natRmfname1name::MFname1pr-LEU2 can1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0).

SGA Analysis: Switcher Method for Starting Strain Construction
We developed a simple 6-step replica-plating method for converting the yeast deletion mutants into a form that enables SGA analysis. This system is simple and takes advantage of the position-confirmed gene deletion mutations constructed by the deletion consortium [E. A. Winzeler et al., Science285, 901 (1999)]. First, the kanMX gene deletion marker of one of the deletion mutants was switched to natMX by transformation with Eco RI-cut p4339, a kanMX to natMX marker-switcher plasmid. To create p4339, we amplified natMX [A. L. Goldstein, J. H. McCusker, Yeast15, 1541 (1999)] from with primers (5'-ACATGGAGGCCCAGAATACCC-3' and 5'-CAGTATAGCGACCAGCATTCAC-3') and introduced the product into pCRII-TOPO (InVitrogen). Second, the resultant natR transformants were mated to Y3084 (MATnamemfname1name::MFname1pr-LEU2 can1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0) and zygotes were isolated by micromanipulation. To create Y3084 (MATnamemfname1name::MFname1pr-LEU2 can1name::MFA1pr-HIS3ura3name0 leu2name0 his3name1 lys2name0), the MFname1 open reading frame of Y3068 was replaced by the LEU2 gene via integration of a PCR product generated with primers (5'-CAAACAAGAAGATTACAAACTATCAATTTCATACACAATATAAACGATTAAAAGAATGTCTGCCCCTAAGAAGATCG-3') and (5'-TACAGTGGGAACAAAGTCGACTTTGTTACATCTACACTGTTGTTATCAGTCGGGCTTAAGCAAGGATTTTCTTAACTTC-3'), which anneal to LEU2 and share homology with the MFname1 locus (in bold), and pRS315 DNA as the template [R. S. Sikorski and P. Hieter, Genetics122, 19 (1989)]. Third, the MATa/name diploids were transferred onto sporulation medium for three days at 22°C. Fourth, MATname meiotic progeny were selected on synthetic medium lacking leucine and arginine but containing canavanine. Fifth, cells were replica plated onto medium containing nourseothricin to select for the deletion mutants. Sixth, cells were replica plated onto medium lacking lysine to identify lys2name0-derivatives.

Deletion Mutant Arrays
A total of 4,672 different yeast deletion mutants, generated by the international deletion consortium [E. A. Winzeler et al., Science285, 901 (1999)], were assembled in our laboratories [Howard Bussey (McGill University)] and also obtained from Research Genetics ( Tables of the deletion mutants contained on the test array shown in Fig. 2A of the manuscript and the large-scale array is provided as supplementary material. Some of the deletion mutants appear to be synthetically lethal with a wild-type control strain; these should include deletion mutants that are slow growing, mutants deleted for genes that are linked to CAN1, deletion mutants defective for histidine biosynthesis, mutants defective for arginine biosynthesis that are synthetically lethal with can1name, deletion mutants defective in mating, deletion mutants that lead to a sporulation defect of the heterozygous diploid, and perhaps others. A table of 437 mutants that appear to be synthetically lethal with a wild-type control strain is provided as supplementary material.

Bioinformatics Analysis
Data from the synthetic lethal interaction network was assembled as a list of yeast gene name pairs. The yeast import tool for the BIND project [Bader et al., Nucleic Acids Res., 29, 242 (2001);] was used to convert the gene name pairs into BIND protein-protein interaction records. This tool integrates yeast information from SGD (, MIPS (, RefSeq (, and the SGD yeast gene registry in order to unambiguously assign any yeast gene name, present in these resources, to an NCBI RefSeq biological sequence. Yeast physical interaction data sets (8429 protein-protein interactions) were imported from YPD, MIPS and previous large-scale genome-wide screens for comparison [P. Uetz et al., Nature, 403, 601 (2000); T. Ito et al., Proc Natl Acad Sci U S A97, 1143-7 (2000); T. Ito et al., Proc Natl Acad Sci U S A98, 4569-74 (2001); B.L. Drees et al., J Cell Biol154, 549 (2001)]. The list of previously known synthetic lethal interactions was created by merging 1142 known synthetic lethal interactions provided by YPD (as of July 29, 2001; with 531 known synthetic lethal interactions from MIPS (as of Aug 7, 2001;, resulting in 1291 unique interactions. The overlap with the genetic interactions presented in Fig. 3 of the manuscript is supplied as supplementary material. The MIPS list was downloaded ( and manually edited to extract synthetic lethal interactions. For network visualization and analysis, BIND can export an arbitrary molecular interaction network as a Pajek network file, which can be viewed with the Pajek program for large network analysis [; D. White, V. Batagelj, A. Mrvar, Computer Review17, 245-274 (1999)]. The format of the Pajek network file can be found on the Pajek web site. Fig. 3 was created with the Pajek program using a Kamada-Kawai automatic layout with subsequent manual alterations to remove node overlap and to cluster the nodes by cellular role. Node coloring by YPD cellular role was done automatically by BIND upon network export using selected cellular role annotation for each protein involved in the network.

A program was written to find yeast genetic or physical interactions in a list where either one or both members of the interaction are of interest. This program was used to identify 72 known physical interactions where both proteins in the interaction correspond to products of genes within the synthetic lethal network presented in Fig. 3 of the manuscript. Homodimers were excluded because SGA can not find interactions of a gene with itself. A map of the 72 protein-protein interactions is included as a figure in the supplementary material. To assess the significance of observing 72 genes whose products interact within this data set, we constructed 1000 random networks that contained 204 genes. On average, 7.8 (~4%; SD = 4.1) genes within the random dataset had products that occurred within the protein-protein interaction data sets. Thus, the synthetic lethal network enrichment for genes whose products interact was highly unlikely to occur by chance. The products of many of the interacting genes occur within pathways probably because the readout of the pathway is required for life in strains carrying the synthetic lethal query mutation.

Two dimensional Hierarchical Agglomerative Clustering and R software ( were used to create supplementary material Fig. 2. The data were represented as a matrix of 0s, 1s and 2s with 201 rows and 8 columns. A cell in the table (row i & column j) with a non-zero entry represents an interaction between array gene i and query gene j. A 1 signifies that the double mutant is synthetic sick and 2 represents that the double mutant is synthetic lethal. The dissimilarity of 2 genes was calculated as the sum of absolute differences over 8 (query gene) attributes. The dissimilarity of 2 screens was calculated as the sum of absolute differences over the 201 (array) genes. The genes and screens were ordered according to their similarity.

Supplemental Figure 1.

Medium version | Full size version

Supplemental Figure 2.

Medium version | Full size version

This file is in Adobe Acrobat PDF format. If you have not installed and configured the Adobe Acrobat Reader on your system, please see Help with Printing for instructions.

Download Supplement Table 3

View Supplemental Table 4

View Supplemental Table 5

View Supplemental Table 6

View Supplemental Table 7

View Supplemental Table 8