Synthetic Lethality

Synthetic LethalitySynthetic Lethality

•Inactivating two interacting pathwaysInactivating two interacting pathwayscauses lethality (or sickness)causes lethality (or sickness)

AA

BB

CC

ProductProduct

XX

YY

ZZ

XViable!Viable!

X Viable!Viable!

Dead!Dead!

Synthetic Lethality

A

B

a

B

XA

bViable Lethal

a

bWild-type Viable

X

• Synthetic Lethality Identifies Functional Relationships

• Large-Scale Synthetic Lethality Analysis Should Generate a Global Map of Functional Relationships between Genes and Pathways

• Gene Conservation = Genetic Network Conservation

A

B

C

EssentialProduct

X

Y

Z

A

B

C

EssentialProduct

X

Y

Z

A

B

C

EssentialProduct

X

Y

Z

A

B

C

EssentialProduct

X

Y

Z

X

Dead

X

DeadDead

X

X X

X

GeneticInteractionNetwork

Similar Patterns of Genetic Interactions Identify Pathways or Complexes

Scenarios That May Give Rise to Synthetic Interaction

• Interpretation depends on context• Each synthetic interaction must be interpreted on a case-by-case

basis (Guarente (1993) TIG, 9:362)

orA B or

regulates

A BA B

etc. etc.

xxxbni1 XMating

MAT MATa

a/wild-type

Sporulation

MATa Haploid Selection(MFA1pr-HIS3)

Double Mutant Selection

Synthetic Gene Array (SGA) Statistics

• 132 query gene mutations were crossed into ~4700 yeast deletion mutants.• Query genes derived from 3 basic functional groups: (1) actin/polarity/secretion,

(2) microtubule/mitosis, and (3) DNA synthesis/repair.• Number of interactions per query varied from 1 to 146 with an average of 36.• (Genes, Genetic Interactions): ~1000 nodes and ~4000 edges.• 17 to 41% false negative rate• False positive rate?• Data quality is good

Making Sense of Genetic Interaction Network

• Correlation with GO annotations

• Hierarchical clustering groups according to their SGA profile– Useful for inferring function of unknown genes

• Correlation with protein-protein interactions?– Only 30/4039 encode physically-interacting proteins

• Statistical properties of genetic interaction network graph

Network of GO Attributes

ClusteringArray

Query

Cell polarity• Actin patches• Endocytosis• Cell wall synthesis• Cell integrity (PKC)

Cell Polarity 20%

Cytokinesis 6%

Cell Wall Maintenance18%Cell Structure

6%

Mitosis16%

Unknown22%

OthersPCL1ELP2ELP3

Vesicular TransportSNC2VPS28YPT6

UnknownBBC1/YJL020c NBP2 TUS1YBL051cYBL062wYDR149cYHR111wYKR047wYLR190wYMR299cYNL119w

MitosisARP1ASE1DYN1DYN2JNM1PAC1NIP100NUM1

Cell Wall MaintenanceBCK1SLT2BNI4CHS3SKT5/CHS4CHS5CHS7FAB1SMI1

Cell StructureATS1PAC11YKE2/GIM1

Cell PolarityBEM1BEM2BEM4BUD6SLA1CLA4ELM1GIN4NAP1SWE1

CytokinesisBNR1CYK3SHS1

bni1: Genome-Wide Synthetic Lethality Screen

Cell Polarity 20%

Cytokinesis 6%

Cell Wall Maintenance18%Cell Structure

6%

Mitosis16%

Unknown22%

OthersPCL1ELP2ELP3

Vesicular TransportSNC2VPS28YPT6

UnknownBBC1/YJL020c NBP2 TUS1YBL051cYBL062wYDR149cYHR111wYKR047wYLR190wYMR299cYNL119w

MitosisASE1ARP1DYN1DYN2JNM1PAC1PAC11NIP100NUM1

Cell Wall MaintenanceBCK1SLT2SMI1CHS3SKT5/CHS4CHS5CHS7BNI4SMI1

Cell StructureATS1PAC11YKE2/GIM1

Cell PolarityBEM1BEM2BEM4BUD6SLA1CLA4ELM1GIN4NAP1SWE1

CytokinesisBNR1CYK3SHS1

bni1: Genome-Wide Synthetic Lethality Screen

DNA RepairASF1HPR5POL32RAD27RAD50SAE2SLX1MMS4/SLX2MUS81/SLX3SLX4WSS1

DNA SynthesisRNR1RRM3YNL218w

MeiosisCSM3

UnknownYBR094w

OthersPUB1RPL24ASWE1SIS2SOD1

sgs1 : Genome-Wide Synthetic Lethality Screen

(24 Interactions)

Chromatin StructureESC2ESC4TOP1

DNA Repair 46%

DNA Synthesis13%

Meiosis4%

Chromatin Structure13%

Cell Polarity4%

Unknown4%

Cell PolarityCell Wall Maintenance Cell StructureMitosisChromosome StructureDNA Synthesis DNA RepairUnknownOthers

8 SGA Screens:291 Interactions204 Genes

Extension of SGA: E-MAP• E-MAP = epistatic miniarray profiles• Quantitative measurement of phenotype (e.g. growth rate)

– Measure both aggravating and alleviating genetic interactions

• Hypomorphic alleles (not null mutations)• Focus on subset of genes• Maya Schuldiner/Jonathan Weissman

Complex A

P

X

X X = NegativePositive=

Complex B

Complex C

Complex X

Complex Y

Complex Z

Organizing Complexes into Pathways Using Genetic Interactions

“RNA World” E-MAP (600 genes)

Positive Genetic Interactions

Negative Genetic Interactions

Positive Genetic Interactions

Negative Genetic Interactions

Proteasome Mutants Suppress Deletions in THP1/SAC3

WT

∆thp1∆thp1 ∆sem1

∆sem1

rpn11-DAmP∆thp1 rpn11-DAmP

∆thp1 rpt6 tsrpt6 ts

Proteasome Mutants Suppress mRNA Export Defects of thp1∆

polyARNA

Nuclei

Merge

WT ∆thp1 ∆thp1∆sem1

polyARNA

Proteasome is Required for Efficient polyA mRNA Export

WT ∆sem1

Complex A

P

X

X X= synthetic lethalityepistatic/

suppressive=

Complex B

Complex C

Complex X

Complex Y

Complex Z

Organizing Complexes into Pathways Using Genetic Interactions

What about essential genes??????

Essential vs. Non-essential Genes in Budding Yeast

Non-Essential Genes (~4800)

Essential Genes (~1050)

3. Conditional point mutants

CREATING MUTANT ALLELES OF ESSENTIAL GENES

1. TET-Promoter Shut-Off Mutants

2. DAmP Alleles

1. TET-Promoter SHUT-Off Strains

-Hughes and colleagues created a library of promoter-shutoff strains comprising nearly two-thirds of all essential genes in yeast (602 genes)

1. TET-Promoter SHUT-Off Strains

-the library was subjected to morphological analysis, size profiling, drug sensitivity screening and microarray expression profiling

1. TET-Promoter SHUT-Off Strains

Cell Morphology

Cell Size

Cdc53

rRNA Processing

1. TET-Promoter SHUT-Off Strains

Gene Expression Analysis

1. TET-Promoter SHUT-Off Strains

Ribosome Biogenesis

Ymr290c, Ykl014c, Yjr041c

Mitochondrial RegulationYol026c

Protein Secretion

Ylr440c

1. Genetic Analsyis using the TET-Promoter SHUT-Off Strains

-30 different mutants X TET-promoter collection

-found many interactions between dissimilar genes

-claimed that there are five times as many “negative” genetic interactions for essential genes when compared to non-essential genes

-however, the cause of this may be due to the fact that the TET strains were very sick (and they were not quantitatively assessing the growth of the double mutant by considering the growth of the two single mutants)

2. DAmP Alleles

(Schuldiner et al., Cell, 2005)

2. DAmP Alleles

3. Point Mutants of Essential Genes

Genetic Profiling of Point Mutants Reveals Insight on Structure-Function

PCNA (Pol30)

-PCNA interacts with CAF-1, a three-subunit protein, to couple DNA replication or DNA repair to nucleosome deposition

-PCNA is important in many aspects of DNA metabolism, including DNA replication and DNA repair

-Two mutants of PCNA (pol30-8 and pol30-79) generated by Stillman and colleagues

Genetic Profiling of Point Mutants Reveals Insight on Structure-Function

PCNA (Pol30)

-PCNA interacts with CAF-1, a three-subunit protein, to couple DNA replication or DNA repair to nucleosome deposition

-PCNA is important in many aspects of DNA metabolism, including DNA replication and DNA repair

-Two mutants of PCNA (pol30-8 and pol30-79) generated by Stillman and colleagues

What is “Chemical Genetics?”

Chemical genetics is the use of exogenous ligands to alter the function of a single gene product within the context of a complex cellular environment.

Find ligands that affect a biological process (forward)Optimize ligands to study protein function (reverse)

Forward Chemical Genetics

• Screening large sets of small molecules

• Goal is target identification

• Those that cause a specific phenotype of interest are used to isolate and identify the protein target

Forward Chemical GeneticsTarget Identification

Plate with cells

Add one compoundper well

Select compound thatproduces phenotypeof interest

Identify proteinTarget(deconvolution)

N

N

Reverse Chemical Genetics

• Screen for compounds that bind to a given protein

• Optimize for selectivity

• Goal is target function and validation

NN

Reverse Chemical GeneticsTarget Validation

Find ligand forprotein of interestOptimize for selectivty

Add ligandto cells

Assay forphenotype

FORWARD Chemical Genetic Studies in Yeast

1. Screening the deletion set for drug sensitivities

2. Comparing mutant profiles to drug profiles

3. Haploinsufficieny analysis

Complex A

P

X

X= synthetic lethality

Complex B

Complex C

Complex X

Complex Y

Complex Z

Organizing Complexes into Pathways Using Genetic Interactions

X= Drug

Alive

Alive

Dead

Alive

Alive

Dead

Synthetic Lethal Interactions Synthetic Chemical Interactions

Deletion Mutants Sensitive to a Particular Drug Shouldbe Synthetically Lethal with the Drug Target

Drug

Drug

1. Screening the deletion set for drug sensitivities

1. Screening the deletion set for drug sensitivities

FORWARD Chemical Genetic Studies in Yeast

1. Screening the deletion set for drug sensitivities

2. Comparing mutant profiles to drug profiles

3. Haploinsufficieny analysis

2. Comparing mutant profiles to drug profiles

Parsons et al., 2004, Nature Biotechnology

1. Clustering of the Drug Profiles:

Camptothecin and Hydroxyurea have a similar mode of action: they both inhibit DNA replication

RFA1RTT105POL30-79POL30-879POL32RAD27RFC5POL30ELG1RFA2PRI1RFC4CDC9TSA1CAMPTOTHECIN (15 g/ml)CAMPTOTHECIN (30 g/ml)

DNA Replication Factors

CAMPTOTHECIN: causes single-stranded DNA nicks and inhibits DNA replication

Also known as : Hycamtin (GlaxoSmithKline) and Camptosar (Pfizer)

-used as an anti-cancer agent

2. Comparison of drug profiles to mutant profiles:

TUB3PAC2CIN1CIN2CIN4BENOMYL (15 g/ml)

Benomyl: a drug that targets microtubules and affects chromosome segregation

CIN1, CIN2, CIN4: genes required for microtubule stability

TUB3: alpha-tubulin

PAC2: tubulin chaperone

2. Comparison of drug profiles to mutant profiles:

FORWARD Chemical Genetic Studies in Yeast

1. Screening the deletion set for drug sensitivities

2. Comparing mutant profiles to drug profiles

3. Haploinsufficieny analysis

3. Haploinsufficieny Analysis

Protein A

P

Protein B

Protein C

Reduced Levels of Protein A

Haploinsufficiency: Drug

Lethality!!!

3. Haploinsufficieny Analysis

TUB1/TUB1 vs. tub1/TUB1

25 ug/ml benomyl 50 ug/ml benomyl

-used a genome-wide pool of tagged heterozygotes to assess the cellular effects of 78 compounds in Saccharomyces cerevisiae

Strategy for Global Haploinsufficiency Analysis Using Microarrays

Comprehensive View of Fitness Profiles for 78 Compounds

No Drug-Specific Fitness Changes

Small Number of Highly Significant Outliers

Widespread Fitness Changes

Molsidomine: potent vasodilator used clinically to treat angina

Erg7: Lanosterol synthase is a highly conserved and essential component of ergosterol biosynthesis

Overexpression of Erg7 results in Resistance to Molsidomine

Identification of Erg7 as the Target for Molsidomine

5-Fluorouracil Targets rRNA Processing

5-Fluorouracil

-one of the most widely used chemotherapeutics for the treatment of solid tumors in cancer patients

-thought to affect DNA synthesis as a competitive inhibitor of thymidylate synthetase

Rrp6, Rrp41, Rrp46, Rrp44: ExosomeMak21, Ssf1, Nop4, Has1: rRNA Processing

The yeast knockout collection

http://www-sequence.stanford.edu/group/yeast_deletion_project/deletions3.html

Using the knockouts for microarrays

A Robust Toolkit for Functional Profiling of the Yeast Genome Pan et al. (2004) Mol Cell 16, 487

Takes advantage of the MATa/ heterozygous diploid collection identifies synthetic lethal interactions via diploid-based synthetic

lethality analysis by microarrays (“dSLAM”)

Uses dSLAM to identify those strains that upon knockout of a query gene, show growth defects

synthetic lethal (the new double mutant = dead) synthetic fitness (the new double mutant = slow growth)

Step 1: Creating the haploid convertible heterozygotes

Important point:This HIS3 gene is only expressed in MATa haploids, not in MAThaploids or MATa/ diploids

So in other words, can select against MATa/ diploids to ensure you’re looking at only haploids later on.

Step 2: Inserting the query mutation

Knockout one copy of your gene of interest (“Your Favorite Gene”) with URA3

Step 3: Make new haploids and select for strains of interest

Sporulate to get new haploids

Select on –his medium to ensure only haploids survive (no diploids)

selects against query mutation so genotype is xxx::KanMX YFG1

selects for query mutation so genotype is xxx::KanMX yfg1::URA3

Reminder about YKO construction

U1 D1

U2 D2

Using common oligos U1 and U2 (or D1 and D2) amplifies the UPTAG (or DNTAG) sequence unique to each of the KOs

Step 4: Prepare genomic DNA and do PCR with common TAG sequences

Step 4: Prepare genomic DNA and do PCR with common TAG sequences

The two different conditions are labeled with two different colors**

The labeled DNA is then incubated with a TAG microarray

**The PCR reactions create a mixture of TAGs (representing all the strains in the pool), since each KO has a unique set of identifier tags (UPTAG and DNTAG) bounded by common oligonucleotides

Evidence this really works – part I

Strains

x-axis y-axis

XXX/xxx::KanMXCAN1/CAN1

XXX/xxx::KanMXCAN1/can1::MFA1pr-HIS3

On average, the intensity is the same before and after 1 copy of the CAN1 gene is knocked out

Evidence this really works – part II

Strainsx-axis y-axis

DIPLOIDSXXX/xxx::KanMX

CAN1/can1::MFA1pr-HIS3

HAPLOIDSXXX or xxx::KanMXcan1::MFA1pr-HIS3

Red spots illustrate that fraction of the strains with KOs in essential genes, so when haploid, not present in pool

Another variation: Drug sensitivity

Another variation: Drug sensitivity

Summary

If you can compare two different conditions and you have a way to stick things to slides, some sort of microarray is possible!

HOW NOT TO LOOK AT INTERACTION DATA!!!!!!!!