Thurs. Mar. 8, 2001 Lecture 2 on Yeast Genetics David Nelson

Yeast Genetic Screens, Two Hybrid, One Hybrid,
Synthetic Lethal and YACs

III. Genetic screens

a) Creating components for use in genetic screens
Genetic screens take advantage of specific properties of proteins and protein
interactions to identify proteins that interact or have redundant activities.
One very common technique is the use of temperature sensitive mutants.
These are called ts mutants and they are made by UV or chemical treatment
of plasmids containing YFG. This may produce only specific types of
mutation (such as transitions) or modification of only one type of base.
Less specific methods involve PCR mutagenesis caused by skewing the
concentration of one nucleotide and by adding Mn2+ to the PCR reaction.
Manganese reduces deoxynucleotide discrimination by the Taq polymerase
and makes it more likely to incorporate an incorrect base. For truely
random mutagenesis, four reactions should be done with a different
deoxynucleotide skewed in each reaction. These can then be pooled at the
end so all types of mutations will be represented. The rate of mutagenesis
should be low (1-2%) so relatively few mutations occur in your gene. The
point is to make ts mutants, not to abolish function. After manipulating
the PCR products into a vector and then into yeast (via E. coli first), the ts
mutants can be detected by growing short streaks of the same yeast colonies
at 18oC, 25oC or 37oC. Yeast that grow at room temperature, but fail to
grow at 37 or 18 are ts mutants. Both cold and warm ts mutants may be
found. The best phenotype is no growth at the non-permissive temperature,
but slow growth compared to the wild type can still be used. The no
growth phenotype is called a tight ts mutant. Slow growth mutants are
called leaky ts mutants. The tight ts mutants are much easier to use in a
screen.

b) Using ts mutants
Once a ts mutant is available, it is possible to clone a complementing gene at
the non-permissive temperature. Usually, a yeast library is transformed into
the yeast host with the ts mutant on a plasmid. The library and the ts mutant
have to be on plasmids with different selectable markers, (like URA3 and
HIS3). The host has to have a knockout of the ts gene, so the ts mutant is
the only copy of the gene present in the host. After transformation, the
yeast are grown at the non-permissive temperature to select for those
colonies that have complemented the ts defect. The plasmids are recoverd
from the yeast and checked by restriction mapping to see if the cloned gene
is different from the original ts gene. It is expected that many of the clones
will be the wild type copy of YFG, but some will probably be different. To
make plasmid recovery easier, strains of E. coli can be used that are
complemented by yeast selectable markers. For example, KC8 cells are
defective in hisB, the homolog of HIS3. This permits HIS3 plasmids to be
selected for on media missing histidine. In this way, only the library
plasmids will be recovered, instead of both. Alternatively, the ts mutant can
be on a CAN1 plasmid that con be chased out by adding canavanine befire plasmid
recovery. Once a clone has been identified that is different from the ts gene,
it can be sequenced. Now that the yeast genome is complete it should only be
necessary to do a single read from both ends of the polylinker to obtain the
complete sequence of the new gene (or genes, since the insert may be
several kb and contain more than one gene).

If you think about this experiment, you will realize that the library has all the
genes that were in the yeast already. So, why would one of these genes
complement the defective ts gene unless it was the same gene? The answer
is that the library can only complement the ts mutant if the complementing
gene is present at higher concentration than normal. These screens are
usually done with YEp vectors (high copy vectors) and the suppressor
genes are called high copy suppressors. If you use YCp vectors (low copy
vectors) then you are only likely to double or triple the concentration of the
plasmid protein product relative to the normal yeast cell. This probably
won't cause complementation of the ts phenotype. High copy suppressors
work by stabilizing a weak interaction between the ts gene product and its
normal partner in a protein-protein interaction. This is done by mass action
and it is much more likely to succeed at high concentrations of the
complementing gene product. This condition will drive the formation of
the protein complex.

A + B -> AB normal situation

A + Bts X ABts complex not favored

(30xA) + Bts -> ABts mass action drives complex formation

c) Complementation cloning from another species.
The same strategy is used as above, except that the library is now derived
from a different species, such as humans. To avoid problems of expression
from foreign promoters, the library may be made as a fusion protein library
with the cDNA from the source being ligated into a strong yeast promoter.
Since this screen is probably looking for homologs of the yeast gene, and
not interacting proteins, low copy vectors may be appropriate.

d) Expression of a foreign gene in yeast.
Assuming that you have a knockout strain prepared, it is possible to express
homologs of YFG from another species. For example, I work with the
yeast gene AAC2 for the ADP/ATP carrier. There are three human
homologs to this protein. They are about 50% sequence identical at the
protein level. If I wanted to express the human genes in yeast, I could
change the promoter to a yeast promoter, alter the codon bias of the first part
of the gene and see if the gene conferred the abilty to grow on glycerol to
my knockout strain.

e) Taking advantage of yeast physiology
Yeast can grow by fermentation, without respiration. In fact, they grow best
by fermentation even in the presence of oxygen. This means that yeast can
tolerate mutations in their mitochondrial proteins that would be lethal in
higher eukaryotes. The typical phenotype of a mutation in a mitochondrial
gene is a glycerol minus phenotype. The yeast cannot grow on a non-
fermentable carbon source. However, the mutant is viable on dextrose (=
glucose). But, these mutants grow slowly on dextrose, so they are called
pet mutants (for petite). It is possible to select for revertant mutations
(gain-of-function mutations) by forcing cells to grow on glycerol. These
mutations can be very informative regarding structure and function of the
protein. Revertant mutations come in two flavors: intragenic and
extragenic. Intragenic mutants are also called second-site mutations,
because they occur in the same gene. Extragenic mutants are in another
gene and they are frequently interpreted as being indicative of a protein
interaction between the two proteins. A suppressor that can supress several
different mutants in a single gene is called a global suppressor. An
extragenic mutant that only suppresses one mutation in a protein but not
another is called an allele specific suppressor. These indicate a highly
specific protein-protein interaction.

IV. Examples of genetic screens

IVa) The two hybrid screen (also called the interaction trap)

Two hybrid systems take advantage of the fact that many eukaryotic transcription
factors have two domains, a DNA binding domain and an activation domain. The
DNA binding domain anchors the transcription factor to the DNA at a sequence
specific site. The activation domain turns on gene expression by interacting with
the transcription complex. The activation domain will work if it is in the vicinity of
the transcription start site. It does not have to be in exactly one specific site on the
DNA, so there is some tolerance where the DNA binding site has to be. The two
domains are modular, and they do not have to be covalently attached to each other
to work. However, they do have to be close to one another. They are often
described as being tethered to one another. The two hybrid screen uses expression
of a reporter gene to assay for an interaction between two proteins that are fused in
frame to the separate transcription factor domains. If they interact, the reporter gene
is turned on.

In the usual in vivo condition, a particular transcription factor controls genes that
recognize its DNA binding domain. However it is possible to switch domains
between transcription factors to create new combinations that do not exist in nature.
An example is replacement of Gal4 DNA binding domain with bacterial LexA
repressor. This hybrid transcription factor will now activate genes that have the
LexA binding site in their upstream sequence.

The birth of this method dates to a paper by Fields and Song in 1989. They
proposed the idea of a two hybrid system for a general protein interaction assay. In
their system, the Gal4 DNA binding domain was fused in frame to a known gene
and it was expressed on a plasmid in yeast. A yeast cDNA library is constructed
with yeast genes fused to the Gal4 activation domain. The library is transformed
into the yeast and the transformants are screened for expression of a reporter gene.
The usual reporter is beta-galactosidase, which gives blue color on X-gal plates.
From the time this was proposed until about 1993, several labs developed systems
that would do this. I have taken the protocols from the lab of Roger Brent at
Harvard. His lab posted a 31 page manuscript on the World Wide Web titled
INTERACTION TRAP CLONING WITH YEAST by Russell L. Finley Jr. and
Roger Brent, to appear in Gene Probes- A Practical Approach published by Oxford
Univ. Press. I have this ms if you want a copy. It has protocols and
troubleshooting advice.

a) Variations on the theme.

DNA binding domains
All systems use either Gal4 or LexA DNA binding domains
These are fused to the known protein called the "bait"

Advantage of Gal4: It targets to the nucleus.
Disadvantage: Cannot use libraries in Gal promoter vectors
because wild type Gal4 must be absent.

Advantage of LexA: No yeast proteins bind to the LexA DNA
binding site. You can use Gal promoter libraries.
Disadvantage: Does not target to nucleus, must overexpress to get it
into the nucleus.

Activation domains (these can be expressed from Gal1(inducible) or
ADH1 or CYC1(constitutive) promoters)

Gal4 (strong activator)
Herpes Simplex Virus protein VP16 (very strong activator)
Bacterial B42 (weaker than Gal4)

Reporter genes
beta-galactosidase (LacZ gene) gives blue colonies
HIS3 these two are selectable markers
LEU2

note: the upstream region of the reporter genes has to have the DNA
binding sequence for the bait fusion protein and the native regulatory
region has to be removed.

b) Embellishments
two reporter genes can be used simultaneously LacZ and one of the
selectable markers like LEU2. Both must have the LexA binding sequence
upstream.

One of the reporters (LEU2) can be integrated into the yeast chromosomes
so that no selection is required to maintain it.

The library fusions to the activation domain can include a nuclear
localization signal and an epitope tag for Western blotting.

The library is on a Gal1 promoter so it can be induced. This way positive
colonies should only be found on induction. They should grow on SC-leu
plates and they should turn blue on X-gal plates.

Elimination of false positives by chasing out the bait carrying plasmid
by a cycloheximide counterselection. Cycloheximide is a protein synthesis
inhibitor. Some yeast strains are resistant to low concentrations of
cycloheximide. A yeast gene called CYH2 makes these strains sensitive
again. If your bait containing plasmid carries the CYH2 gene, then yeast
plated on cycloheximide will die unless they lose this plasmid.
This is called a counterselection and it is the same principle as using
5-FOA with a URA3 plasmid or canavanine with a CAN1 plasmid.
Positive colonies from the 2-hybrid screen should be treated with
cycloheximde to chase out the bait plasmid and then rechecked for
a positive signal. If it is still positive, it is a false positive and can be
eliminated.

c) More about the bait
Usually the bait fusion is expressed from a strong promoter like the ADH1
promoter. If the bait does not get into the nucleus, a nuclear localization
signal can be added.

The sensitivity of the assay can be increased by adding more LexA binding
sites upstream of the reporter. Three LexA sites can be added instead of
one.

Problem: Some bait fusions will activate the reporter without the activation
domain present (before transformation with the library). This may indicate
some ability of the bait to act like an activation domain on its own. To get
around this, a less sensitive reporter can be used (only one LexA binding
site) or some of the bait can be removed. Activation domains are often
acidic, so an acidic region of the bait could be deleted.

d) Quantitation
LacZ activation is dependent on the strength of the interaction between the
bait and the unkown partner fused to the activation domain. the stronger the
interaction the more beta-galactosidase will be made. Sensitive assays exist
to measure this activity, so the strength of the interaction can be evaluated.
It does not take much of the LEU2 gene product to survive on SC-leu
media, so the LEU2 gene is a very sensitive indicator of activation. Its
response tends to be plus or minus. The LacZ gene is a better gauge of the
strength of the interaction.

e) A typical plasmid for the library construction
A high copy plasmid with a TRP1 selectable marker can be used.The library
is under control of the Gal1 promoter. The 1st 9 codons will be SV40 virus
large T antigen (for nuclear localization). The next 87 codons are for the
B42 activation domain. The next 9 codons are an epitope tag for the HA
(haemagglutinin) protein. This is followed by EcoRI and XhoI sites for
cloning in the library fragments.

f) Most successful strategy
Do a two step transformation and selection protocol

1) transform in the library but only select for Trp1 (on the library plasmid)
Save a mixture of the transformants in a frozen stock.

2) Plate aliquots of step one onto Gal , SC-leu plates. This induces cDNA
fusion proteins to be made and it selects for interacting proteins.
Doing both steps in one may cause failure.

g) Interaction mating
Testing 5 baits against 10 potential partners (sometimes called the prey) can
be labor intensive. The 50 different plasmid combinations would require 50
different transformations. To avoid all this labor, it has been suggested that
the bait plasmids could be transformed into a MATa host strain and the prey
plasmids could be placed in a MAT alpha strain. Then the 50 crosses could
be made in a few minutes, avoiding 35 yeast transformations. This is called
interaction mating. It has been taken to extreme lengths by crossing whole
viral genomes with each other (Bacteriophage T7). Something like 25
protein interactions were discovered this way.

IVb) The One hybrid screen.

The one hybrid screen skips the protein interaction step required to activate the
reporter gene in the two hybrid screen. Instead, it relies on a DNA-protein interaction to
turn on the reporter gene. The object of this kind of screen is a DNA binding protein, with
a known DNA binding sequence. The DNA sequence is called a target element.

To do a one hybrid screen you must make a modified reporter gene with the DNA sequence
you want (target element) incorporated upstream of the reporter gene. Once you have made
the construction with the target element in front of the promoter, the plasmid is linearized
and integrated into the yeast genome. The resulting strain is called a reporter strain. This
strain is then transformed with a library of candidate cDNAs that have been fused to the
activation domain for the reporter gene. If you are using the GAL4 system you would have
a GAL4 activation domain fusion library to activate the reporter, that can be beta
galactosidase, or HIS3 or both. The advantage of using both reporters is reduction of false
positives. In using HIS3 as the reporter gene, colonies will only grow if the HIS3 gene is
activated, but there is some leakyness in this method. To reduce false positives, you can
add a small amount of 3-AT (3-amino-1,2,3-triazole) that is an inhibitor of HIS3. The
colonies that grow under these conditions will have a significant expression of the HIS3
gene and are less likely to be false positives.

IVc) The synthetic lethal screen (a sectoring assay)

Proteins are often redundant in their function. If one protein does not work because
of a mutation (or a gene disruption by a researcher) another protein may fill in and
do the same job. This leads to no recognizable phenotype when that gene is
disrupted or deleted.

If you have cloned a gene in yeast by a genetic screen, but the gene does not have a
phenotype when it is disrupted, two interpretations can be made. 1) the protein is
not important and the cell can do without it. 2) The protein is redundant and it has a
backup. If the cell has a backup or redundant partner, how can you discover what
the other gene is? You can do what is called a synthetic lethal screen.

The idea behind this screen is that a protein not normally essential for viability may
become essential if its redundant partner is mutated. This assumes that only one
other protein is backing it up.

To do this screen, you first make a deletion strain of your cloned gene. Then you
put this gene back in on a plasmid with an ADE1 selectable marker. The host strain
is an ade1 mutant that causes buildup of a red pigment when adenine concentrations
are low. The red pigment is an intermediate in adenine biosynthesis that builds up
when the pathway is turned on. High adenine will feedback inhibit the pathway
and the colonies will be white. An ADE1 plasmid will make the missing enzyme
and the pigment will not accumulate, so the colonies are white.

If there is no selection for the plasmid with the ADE1 gene on it, the plasmid will be
lost at 1-2% per generation. This causes the yeast to turn from white to pink. If a
colony has already started forming, a sector of the white colony will turn pink.
These are the cells that grew up after one cell lost the ADE1 plasmid. The ability to
form pink sectors is the basis of the synthetic lethal screen.

To do this screen, you must mutate the knockout strain with the wild type gene on
an ADE1 plasmid. Chemical mutagenesis or UV light will work. After this is
done, colonies are identified that no longer can form pink sectors. This is assumed
to be caused by mutation of the backup gene. Now the plasmid borne gene is
essential and the cell cannot lose it, so the colonies do not sector. The cloned gene
has developed a synthetic lethal phenotype because the backup gene is no longer
functional.

Now you transform a yeast genomic library into the mutagenized strain and see if
any of the colonies are able to sector once more. This new gene would be the
complement of the redundant gene. Once you have sequenced it you must do a
knockout of the new gene to see if it has a phenotype. Probably it will not, since
the first gene will be redundant for it. However, when you make a cross of the two
knockout strains, the result should be lethal if your hypothesis was correct. This is
a tedious way to identify redundant genes. It takes many months and thousands of
plates and lots of patience.

V. YACs (yeast artificial chromosomes)

In 1987 Maynard Olson's lab described yeast artificial chromosomes for use in
cloning large fragments of DNA from humans and other eukaryotes. Burke, D.T.,
Carle, G.F. and Olson,M.V. Science 236, 806-812 (1987). The YACs contain
three essential parts.

1) a centromere
2) an ARS (autonomously replicating sequence)
3) two telomeres

The YACs have to be 20-50kb long to be stable. Larger YACs are more stable than
smaller ones. The YACs fill a gap in the size range of DNA analysis. Cosmid
vectors work up to about 50kb. Chromosomal banding methods and genetic
mapping have a lower resolution of about 500kb to a Mb. The YACS can be up to
2.5 Mb, allowing physical maps to be constructed that are on the same scale as
genetic maps. 10,000 YACs can cover the human genome, while 100,000 cosmids
would be required.

a) Vectors for YAC cloning
The 1987 paper in Science describes pYAC2 and three variants, pYAC3, 4
and 5. The vector is a small E. coli yeast shuttle vector, based on pBR322
with amp resistance. CEN4 is the centromere from yeast ChrIV. TRP1 and
ARS1 are adjacent to the centromere and they provide the ARS and a
selectable marker required by the YAC in yeast. The TRP1 gene is a marker
for the short arm of the YAC. URA3 is a marker for the long arm (where
the DNA insert will be). These markers are used so clones can be selected
that have a left and a right telomere, which they need to be viable in yeast.
The telomeres in the vector are not true telomeres. They are in a closed
circular plasmid, and true telomers are found on the ends of linear
chromosomes. How are telomeres generated from plasmid sequences?
This is done by cleaving a sequence that leaves a telomere seeding sequence
on the end of the DNA. This telomere seed is recognized by the cell as a
telomere that needs to be repaired and the cell completes the ends. The two
telomere sequences bracket a HIS3 gene on the plasmid. The HIS3 gene is
removed by BamH1 digestion to expose the telomere seed sequences.

In addition to all these required features, the cloning site is located in a gene
called sup4-o. Sup4-o is a suppressor tRNAtyr gene that reads an ochre
stop codon UAA as a tyrosine. This gene has a natural SmaI site (blunt
cutter), that is used to clone the large DNA fragments. If a DNA fragment is
inserted in this site the sup4-o gene is disrupted and cannot suppress the
UAA stop codon anymore. When an empty YAC is transformed into a
yeast host that carries a modified Ade2 gene with a UAA stop codon in its
sequence, the sup4 gene product suppresses the stop codon and the yeast
make a functional ADE2 gene product. The cells are white. If a DNA insert
has disrupted the SUP4 gene, the cells don't make ADE2 and the cells are
pink. This is a simple color test to see if the YAC carries an insert in the
cloning site. It is similar to blue white screening with beta-galactosidase.

The Sup4 gene also has a short intron of 14bp. Since it was not possible to
modify the tRNA gene sequence to add more restriction sites, the
intron was chosen as a site to introduce more sites. First, an Sna BI site
(blunt cutter) was added in this region to make pYAC3. Then a series of
short linkers were added in at this site to add an EcoRI site (pYAC4) or a
NotI site (pYAC5). Disruption of the intron by a large DNA fragment also
stops the Sup4 gene from making a product. Since these early vectors,
there have been some improvements. pYAC55 has replaced pYAC5 since
some of the pYAC5 vectors were defective. There is no difference except
the defect is gone. pYAC-RC has a polylinker with six restriction sites.
pYAC4-neo adds G418 resistance for use in mammalian cells.

b) Problems
1) shearing of DNA. The DNA inserts are very large and have to be
handled very gently. Transformation protocols use the protoplast method
rather than the Li Acetate method. The worst part about this method is that
the cells are plated in a top agar so colonies form in solid agar and cannot be
replica plated.

2) Chimeras. Because of the way the DNA inserts are added to the vector
sequences, three different fragments must be ligated together. Sometimes
more than three fragments will be ligated. This will cause DNA from
different regions to be joined in a chimera. Such chimeric DNA sequences
make DNA mapping difficult.

3) Rearrangements. Large DNA fragments in the YACS can undergo
rearrangements that scramble the DNA sequence. The human genome
project was done using BACs (Bacterial Artificial Chromosomes) that
are more stable.

c) Applications
1) expression of very large genes. It is possible to introduce YAC DNA
into mammalian cells by spheroplast fusion. The new resistance marker to
G418 allows the DNA to be selected for. In this way very large genes of
several hundred kb can be expressed in mammalian cells.

2) genome projects(mapping, cloning, sequencing) An ordered YAC library
can cover the whole human genome in a relatively small number of clones.
These ordered libraries are being constructed by mapping the overlaps of the
YACs to produce a contig for each human chromosome.

3) gene hunts for disease genes. Genetic mapping of disease genes can
identify the chromosomal region of a gene. If YACs for that region are
available, then the gene hunt can focus on the YAC DNA.

Yeast Genetic Screens, Two Hybrid, One Hybrid, Synthetic Lethal and YACs

Yeast Genetic Screens, Two Hybrid, One Hybrid,
Synthetic Lethal and YACs