Overview
The interaction trap is a method for cloning cDNAs encoding proteins that~interact with a known protein. The method also provides a simple and efficient genetic assay for protein/protein~ interactions. Each interacting partner is epitope tagged~so that the interactions can also be detected physically. Moreover, the individual components have been designed so that the system provides a number of internal controls and gimmicks that simplify its use. The interaction trap was developed to aid our search for Gl specific cell cycle regulatory proteins in higher eukaryotes, primarily by Jeno Gyuris and Erica Golemis, with contributions from other members of the Brent lab. It evolved from attempts in the lab (since 1986) to use LexA fusion proteins and LexA responsive reporter genes to find interacting proteins encoded by cDNA l~libraries. The trap owes a great debt to the work of Fields and Song, who proposed in~l989 that library encoded proteins that interacted with DNA bound fusion proteins could be detected if the library encoded proteins all carried transcription activation domains. This idea was critical, and the interaction trap can be considered a particular version of the two-hybrid system these workers proposed. Numerous other researchers in other labs have also developed useful implementations of this theme, many of which also Contain useful refinements.. For the technology described here, we ask you to take note of the names of ~he postdocs and students who created, over the past 7 years, the various components and ideas that made it possible.
How IT Works
In the interaction trap, the protein of interest, referred
to as the "bait", is expressed as a LexA fusion in a
"selection strain"', a yeast strain containing LexA
binding sites ('operators") upstream of a selectable marker
gene, LEU2.
Expression of a cDNA library with the cDNA-encoded proteins fused
to a transcription activation domain allows selection of those
that interact with the bait because they activate transcription
of LEU2. Yeast containing cDNAs that encode bait interactors grow
on media lacking leucine. The yeast strains used also contain
one member of a series of different LexA operator-lacZ reporters.
Thus, growth of a cell that contains a library plasmid into a
colony in the absence of leucine indicates an interaction, and
the amount of betagalactosidase activity produced points to the
strength of the interaction.
The following descriptions of the interaction trap are based on
our experience with several successful hunts for interacting proteins.
We are always tinkering with the interaction technology in response
to results from here and-
from the outside world. We will attempt to update this document
frequently with significant improvements as these are made.
Please contact us with any comments or questions about this information. Please also feel free to use and distribute these documents as you wish, but we do ask that you do not change the wording, and keep the names of the authors - attached to ~the information that they contributed..
Russ Finley
Roger Brent
Department of Genetics
Harvard Medical School and
Department of Molecular Biology
Massachusetts General Hospital
50 Blossom Street
Boston, Massachusetts 02114
USA
finley@ frodo . mgh harvard. edu
brent@frodo .mgh.harvard.edu
YEAST STRAINS
The yeast strains we now use for the interaction trap were
made by Erica Golemis. The most important feature of these strains
is that they contain a selectable gene, LEU2, which has upstream
LexA binding sites (LexA operators; LexAop) in place of its normal
upstream activating sequences (UAS). The LeXAop-LEU2 gene replaces
the chromosomal LEU2 gene and therefore is stable (requires no
selection to maintain it). The LexAop-LEU2 gene is normally not
transcribed and the yeast are auxotrophic for leucine. LexA fusion
proteins that activate transcriptional.,operators, activate LEU2
transcription, and allow these strains to grow in the absence
of leucine. LexA fusion proteins that do not activate transcription
can be used as baits for the trap. Use of transcriptionally inert
bait proteins allows one to express in the same strain cDNA-encoded
proteins fused to a transcription activation domain, and select
those cDNAencoded proteins that interact with your bait by their
ability to cause leucine prototrophy. Because the cDNA fusion
is expressed from a galactose-inducible promoter (a GALl derivative,
see LIBPARIES), it is possible to quickly determine that the leucine
prototrophy is dependent upon cDNA expression; cells containing
true interactors will grow on media lacking leucine only if it
containS galactose but not when it contains glucose. During an
interactor hunt we include a second reporter gene in the Strain
to provide an independent measure of interaction. The second reporter
(see REPORTERS) is a plasmid-borne lacZ gene with upstream LexA
operators. Yeast containing a cDNA that encodes a true interactor
will show galactose dependence of both phenotypes, they will exhibit
galactose-dependent growth in the absence of leucine and galactose-dependent
blue color on Xgal indicator plates.
The strains are auxotrophic for several other markers which respectively allow selection of the bait plasmid (H153), the lacZ reporter plasmid (URA3) and the library plasmid (TRPl)
The prototype strain is EGY48 (alphaA, his3,trp1, ura3, 6LexAop LEU2) . EGY48 has 3 high affinity colEl LexA operators upstream of the LEU2 gene. Each operator consists of two overlapping LexA binding sites, each of which can bind a dimer of LexA. This makes for a very sensitive reporter of transcription activation by LexA fusions or by proteins interacting with them. However, baits that activate the LEU2 gene in EGY48 cannot be used for the interaction trap. For this reason Erica has also made strains with fewer LexA operators; these allow use of LexA fusions that marginally activate the LEU2 gene in EGY48 as baits.
Below is a detailed description of how Erica made EGY48 and
its derivatives.
LexAoperator-LEU2 selection strains Erica A. Golemis
Construction
A Hindiii cassette encoding the URA3 gene was inserted into a
derivative of the plasmid pHR33 (a gift of R. Rothstein) to create
the yeast integrating plasmid pXLEU2. 1,2, or 3 copies of a BamHI-ended
double stranded 42-mer that contained the "overlapping"
double LexA operator found upstream of tne colecin El gene (see
below) were inserted into a unique BgIII site -130 bp upstream
of the LEU2 translation start site, to generate plasmids lLexLEU2,
2LexLEU2, and 3LexLEU2.
The strain U457 (R. Rothstein) (MATa 5UP53-a ade2-l canl- 100
ura3-1 trpl-l [phi+]) was used as recipient for the LexAoperator-LEU2
plasmids. The trpl-l mutation is an amber, which is suppressed
by the 5UP53-a gene, so in fact U457 is "TRP+" for growth.
The chromosomal array of genes in this strain is Tyl element-5UP53-LEU2.
Integrating plasmids were cut with Cla I within LEU2 coding sequence,
and used to transform U457 to URA3+. ura3 revertants were selected
by their ability to form colonies on medium that contained 5-
Fluoroorotic acid. Ura3~ trpl~ leu2~ yeast derive from recombination
events that resolve the integrated plasmid by crossover between
TyT sequences present on XLEU2 variants and the chromosomal Tyl
element, resulting in loss of the 5UP53-a gene and substitution
of LexA operator sequences for the LEU2 UAS. ura3- trpl- leu2-
colonies containing 0,1,2, or 3 colEl operators upstream of the
single chromosomal LEU2 gene were isolated. EGYl2 has no LexA
operators, EGYl7 and EGYl8 have one, EGY23 and EGY24 have two,
and EGY38 has three.
Strains
EGYl2, EGY17, EGY22 and EGY38 were made his3- by mating these
strains to the strain GGIOO-14D (MATalpha his3 trp1 pho5); selecting
for LEU+, HIS+ diploids: sporulating, and selecting for random
spore products that were leu2~, ura3~, trpl~, his3, and GAL+.
EGY48, EGYl94, and EGYl188 are derivatives of EGY38), EGY22, and
EGYl7, respectively. EGY48 is most commonly used for the interactive
library screen described elsewhere. EGY40 is an EGYl2 derivative
used as negative control.
MAKING BAITS
LexA fusions proteins ("baits") are typically expressed
from
MAKING BAITS LexA fusions proteins ("baits") are typically
expressed from pEG2O2. Protein coding sequences can be inserted
into this plasmid in a polylinker downstream of the region encoding
full length LexA (amino acids 1 to 202, including the DNA- binding
domain and the dimerization domain). Although LexA does not contain
a functional yeast nuclear localization signal, it and most fusions
to it will be equally partitioned between the cytoplasm and nucleus
(Silver, et al. (1986) MCB 6:4763 4766) (see TESTING BAITS) .
Thus, most baits will occupy LexA operators in the yeast nucleus
(Golemis and Brent, 1992 MCB 12:3006-3014). A small minority of
baits are excluded from the nucleus (for-example, LexA-Oskar;
Breitwieser and Ephrussi, personal communication) . For these
rare baits, we have available a close relative of pEG2O2, a gift
of Bert Vogelstein, which carries the SV40 T nuclear localization
signal between LexA and the polylinker. We also have a pEG2O2-related
plasmid for making fusions to the N- terminus of LexA, a gift
of Manuel Sainz and Vicki Chandler. This vector can be used in
those [rare.] cases where it is suspected that the N-terminal
region of a protein may be importantfor interaction. - LexA Fusion
Expression Plasmid, pEG2O2.
Erica A. Golemis
pEG202 is a derivative of the plasmid LexA2O2+?L (Ruden et
al., Nature 350:250-252 (1991)) which increases the number of
unique polylinker sites available for cloning. To create this
plasmid, LexA202+PL was cut at the unique SalI cloning site downstream
of LexA in the polylinker, and a 22mer with the sequence
5' TCGACCATGGCGGCCGCTCGAG
GGTACCGCCGGCGAGCTCAGCT 5 I
was inserted. This oligo thus recreates flanking SalI sites, and adds novel unique. sites for NcoI, NotI, and XhoI. The total pEG2O2 polylinker sequence reads in frame from the last codon of LexA as follows:
CTG GAA TTC CCG GGG ATC CGT CGA CCA TGG CGG CCG CTG GAG TCG ACC
LexA ECORI Smal BAMHI SALI NCOI NOTI XHOI SALI
TGC AGC C.
PstI
All sites indicated in capital letters are usable for inserting fusion sequences (PstI and Smal are not unique) . This plasmid contains the 2um origin and a HIS3 selectable marker and the E.coli amp resistance gene. Expression of the LexA-fusion cassette is from the. strong constitutive ADH1 promoter. Primers
The colEl operator
LexA operator-containg oligonucleotide. Contains very high affinity
colE1 sites which appear to be bound by two LexA dimers. underlined
bases' are important for protein contacts (Ebina et al JBC 258,
13258-13261 (1983); Kamens et al., MCB, 10, 2840-2847(1990)).
5' GATCCTGCTGTATATAAAACCAGTGGTTATATGTACAGTACG 3'
3' GACGACATATATTTTGGTCACCAATATACATGTCATGCCTAG 5'
Erica Golemis March, 1990
Department of Molecular Biology, MGH and
Department of Genetics, Harvard Medical School
50 Blossom Street
Boston, Massachusetts 02114
After August 1993
Fox Chase Cancer Research Center Philadelphia, Pennsylvania
[email protected] Information revised September 1993
Erica A. Golemis
Reporters
Selection for LEU+ colonies in an interactor hunt results in a certain number of false positives due to cis or transacting yeast mutations that activate the LEU2 promoter (see EXPECTED RESULTS). Most of these can be identified because they grow on glucose plates lacking leucine and therefore do not depend on expression of the activator containing protein from the library plasmid which is driven by the GALl promoter (see LIBRARIES) . However, the selection strains for the trap also incorparate a second reporter, which has LexA operators upstream of a lacZ gene (see below), to further diminish the frequency of these apparent positives. Most yeast that contain library encoded proteins that interact with the bait will be both galactose dependent LEU+ and galactose-dependent blue on Xgal indicator plates.
The LexAop-lacZ reporters are not as sensitive as the Lexop-
LEU2 reporter, so it is possible for a weak interactor not to
result in blue yeast on Xgal plates. To minimize this possibility,
for most hunts we now recommend the use of the most sensitive
lacZ reporter plasmid available, usually-
(pSHl8-34} rather than the less sensitive ones used previously
(pJK1O3, and pRB1840) (Gyuris et al., submitted; Zervos et al.,
January 1993 Cell 72:223-232 ) . Note that these less sensitive
lacZ reporters can be quite useful:
activation of their transcription by a particular interacting
bait/prey pair is a function of the the dissociation constant
for that interaction (Golemis and Brent, in preparation) and demanding
activation of weak reporters is a good way to detect only strong
interactions.
pSHl8 series - LexAop-lacZ reporter genes
Steve Hanes-
These are reporter plasmids for LexA fusion Proteins. They
contain multiple LexA operators upstream an easily assayable reporter
gene (GALl-LacZ) They are all URA3+ and have 2u replicators. In
E. coli, they all confer amp resistance.
I made these by inserting a 78-bp oligonucleotide* into the XhoI ~ite of plasmid pLRl(delta)l (West et al., MCB 4, 2467 (1984)). The XhoI sites are reconstituted. Insertion site is at -167 from the transcription start site of GALlLacZ. Inserts are in the 5'-> 3' orientation. Operators are the high-affinity "overlapping" sequence found upstream of coIEI'(Ebina et al., JBC 258 13258). Each oligonucleotide contains two of these s~nces (ie, 4 binding sites for LexA dimers, or 8 operator half sites)
pSHl8-8: One intact insert. 2 LexA operators.
pSHl8-3: One and a 'half 1 inserts (rearranged in vivo). 3 LexA operators.
pSHl8-34: Two inserts. 4 LexA operators.
p5H18-34(delta)spe: Integrating derivative of pSHl8-34
Comments: Because of their large numbers of high affinity operators,
these plasmids are currently the most sensitive
Lexop-lacZ reporters. Yeast bearing them have undetectable levels
of beta-galactosidase activity in the absense of activator proteins.
PSHl8-34(delta)spe is a gift of Pam Silver who is now at the Dana
Farber Cancer Institute. It lacks the 2 um replicator and can
be directed to the URA3 locus for chromosomal integration by cutting
with Apal.
Also available is JKlO3 (made by Joanne Kamens, Brent
lab), which has a single overlapping colEl type operator, and
the old faithful, p1840 which has a symmetrically altered version
of the lower affinity recA operator (1 LexA op; Brent and Ptashne,
Cell 40., 729. (1985)). For all of these LexA op ; reporter plasmids
appropriate negative control is PLR which gives no background
in yeast.
*Sequence of the oligonucleotide Each oIigonucleotide contains 2 colEl operators or 4 binding sites for LexA dimers.
top strand
5'tcgactgctgtatataaaaccagtggttatatgtacagtactgctgtatataaaaccagtgg
ttatatgacagtacg 3'
bottom strand
3'gacgacatatattttggtcaccaatatacatgtcatgacgacatatattttggtcaccaata
tacatgtatgcagct 5'
These anneal to give double stranded DNA with 4bp XhoI 5' sticky
ends
(TCGA)
Steve Hanes
Department of Molecular Biology
Massachu~etts General Hospital
Boston, MA 02114
after september 1, 1993
Axelrod Institute
Wadsworth Laboratories
New York State Department of Health
Empire State Plaza
Albany, New York 12201
518-474-2821
[email protected]
Information revised September 1993
Roger Brent
brent@opal .mgh.harvard. edu
Testing baits
Testing baits
The bait plasmid Should be tested in yeast to show that it directs
the synthesis of the fusion protein, and that the bait enters
the nucleus, and binds LexA operators.
To be useful in an interactor hunt, your bait must not activate
transcription by itself. pSHl8-34 (see REPORTERS) and yeast strain
EGY48 (see STRAINS) can be used to test whether your bait activates
transcription. Introduce pSHl8-34 and a bait plasmid into EGY48
by transformation, and patch transformed colonies onto Xgal indicator
plates and leucine minus plates. As described in the accompanying
handout from Steve Hanes (see REPORTERS), pSHl8-34 is a U~RA3+,
2um plasmid containing the GAL1 promoter fused to lacZ, in which
the GALl enhancer-like Upstream Activating Sequence (UASG) has
been replaced with bindinq sites for 8 LexA dimers Baits that
activate transcription cause expression of beta-galactosidase
from this reporter. You should also ensure that your bait does
not activate the LexAOp-LEU2 gene in EGY48 (see YEAST STRAINS).
To do this, test whether your bait plasmid allows growth of EGY48
in the absence of leucine.
_
_
Baits that do not activate transcription should be tested to see
whether they enter the nucleus and bind LexA operators. This can
be tested by using the bait plasmid with another reporter, PJKlOl,
in a repression or "blocking" assay (Brent and Ptashne,
1984, Nature, 312:612-615), which detects DNA binding by transcriptionally
inert LexA fusion proteins. For this assay, EGY48 is cotransformed
with the bait plasmid and PJK1O1. PJKlOl is a URA3+, 2um plasmid
similar to sH18-34, except that it contains most of the GAL1 upstream
activating sequence (UASG) . It also contains, in between UASG
and the GAL1 transcription start, a colEl derived LexA operator,
which provides high affinity binding sites for two LexA-bait protein
dimers. Yeast harboring pJKlOl will have significant_beta galactosidase
activity when grown on galctose medium LexA fusions that enter_the
nucleus and bind the operators; but do not activate transcription
repressing this beta-galactosidase activity from 2 to 20 fold.
We take even 2 fold repression to indicate >50% operator occupancy
by the bait.
Two H153+ control plasmids are commonly used for the activation
assay and repression assay One, pSHl7-4 expresses a LexA-GAL4
fusion that activates transcription. The other, PRFHM1, expresses
LexA fused to a transcriptionally inert fragment of the Drosophila
Bicoid product (pEG2O2 does not serve as a good negative control
because the part encoded by the polylinker through to the first
stop codon activates transcription a bit) pRFHMl can also be used
as a control for the repression assay; the LexA-bicoid fusion
represses the activity of pJKlOl by 2 to 5- fold.
Russ Finley
revised September 1993
LIBRARIES
pJG4-5 is the yeast vector used for constructing interaction libraries. It was made by Jeno Gyuris in the Brent lab. It is a 2um plasmid containing the selectable TRP1 marker and the E.coli amp resistance gene. cDNAs are inserted into a unique ECORI and/or an adjacent unique XhoI site) In-frame cDNA encoded proteins will be expressed with a fusion moiety at their amino terminal end. The fusion moiety consists of an SV40 nuclear localization signal, followed by an acidic transcription activation domain called B42 (Ma and Ptashne, 1987), and a 9 amino acid hemaglutinin epitope tag. The promoter driving the expression of the fusion moiety is the yeast galactose inducible, glucose-repressible GAL promoter. Transcription is terminated by the yeast ADHl terminator which follows the EcoRI and XhoI sites. See SEQUENCES for making sequencing primers.
Below are descriptions of a Hela library made by Jeno Gyuris, Drosophila libraries made by Russ Finley, a yeast library made by Paul Watt, and a serum-starved WI-38 cell library made by Jeno Gyuris and Claude Sardet. -
Hela cell acid fusion library - Jeno Gyuris
I isolated RNA from serum grown, proliferating Hela cells that were grown on plates to 70% confluence. I extracted the RNA by lysing the cells with guanidium followed by treatment with phenol chloroform, and purified polyA+ mRNA on an oligodT cellulose column. I made the cDNA using a scheme similar to the variation Huse and Hansen (Strategene Strategies, 1, 1-3, 1988) described of the Gubler and Hoffman technique (Gene, 25, 263 269, 1983)
That is, I made the first strand of cDNA using a linker- primer
(JG33) that contained , 5' to 3', an 18nt poly dT tract (to hybridize
to the mRNA's polyA tail), an XhoI site, and a 25 nt sequence
to protect XhoI site. For first strand synthesis, I used Supercript,
from BRL, which is an recombinant RNAseH defective verion of the
Moloney virus reverse transcriptase, and I used 5medCTP instead
of dCTP, so that internal XhoI sites in the cDNA would not be
cleaved by XhoI. For second strand synthesis, I treated the mRNA/cDNA
hybrid with RNAseH and E. coli DNA polymerase
I. I made the resulting ends flush by sequential treatment with
Klenow, Mung 'Bean exonuclease, and Klenow, then ligated EcoRI
adaptors
5' AATTCGGCACGAGGCG 3'(JG3l)
3' GCCGTGCTCCGC 5'(JG32)
onto the ends, and digested the cDNA with XhoI. I size- fractionated
this DNA on a 5-20% KoAc gradient, collected fractions that contained
>700 bp fragments, and ligated these into EcoRl- and XhoI-cut
PJG4-5, a 2u, TRP1+, GALl promoter vector similar to pJG7-
4 except that it confers amp instead of kan and has the pucl9
origin of replication instead of the pMBl origin.
I collected 9.6 X 106 primary transformants by scraping LB amp plates. I pooled these and grew them in LB medium overnight, and prepared plasmid DNA. 90% of the library members contain a cDNA insert whose size ranges between lkb-2kb. Western blots of individual yeast transformants using ascites supernatant from the anti hemagluttinin monoclonal suggest that between 1/4 and 1/3 of the members express fusion proteins.
Jeno Gyuris October 16, 1991
Present Address:
Mitotix Incorporated
One Kendall Square
Building 600
Cambridge, Massachusetts 02139
[email protected]
Drosophila libraries - Russ Finley'
Drosophila embryo acid fusion library (RFLY1) November 1991
Unidirectional cDNA made from RNA from 0-12 hour wild type
embryos and inserted into pJG4-5. 4.2 x 10e6 individual E.coli
transformants were recovered. Over 90% of the plasmids of this
library contain cDNA inserts whose sizes range from 0.5 to 2.5
kb (average size about 1 kb)
Drosophila embryo epitope' tagged library (RFLY2) June 1993
Unidirectional cDNA made- from RNA from 0-12 hour wild type embryos and inserted into pJG4-6. 3.2 x 10e6 individual E.coli transformants were recovered. Over 91% of the plasmids of this library contain cDNA inserts whose sizes range from 0.5 to 2.0 kb (average size about lkb)
Drosophila ovary acid fusion library (RFLY3) June 1993
Unidirectional cDNA was made from wild type Drosophila melanogaster 'ovary RNA (isolated and poly(A)+ selected by Gerardo Jimenez in David Ish-Horowicz's lab) and inserted into pJG4-5. 3.2 x 10e6 individual E.coli transformants recovered. Over 87% of the plasmids of this library contain cDNA inserts whose sizes range from 0.3 to 1.5 kb ,'average size about BOObp)
Drosophila disc acid fusion library (RFLY5) June 1993
Unidirectional cDNA was made from poly(A)+ selected disc RNA (provided
by Janice Fisher Vise in Ruth Lehman's lab) and inserted into
pJG4-5. 4.0 x 10e7 independent E.coli transformants were recovered.
Over 92% of the plasmids of this library (RFLY5) contain cDNA
inserts whose sizes range from 0.3 to 2.1 kb (average size about
9OObp).
Russ Finley
(617) 726-5956 (phone)
(617) 726-6893 (fax)
fin ley@ frodo . mgh harvard. edu
Yeast interaction library pJG4-5 library
June 1993
This was made by Paul Watt, a postdoc in Jim Wang's lab, who requests that his name be included as an author in the first paper resulting from its use. According to Paul....
50 ug of genomic DNA from 5288c (MAT~ SUC2 mal gal2 CUPl hapl), of apparent size >100~b, was split into two 25ug portions. One portion was partially digested with Alul, the other with HaeIII, to yield an average (median) size of lOOObp. Each of these reactions was than treated with EcoRl methylase and the extent of methylation protection confirmed by digestion of small aliquots with EcoRl and subsequent visualization on agarose gels. An equimolar mixture of 3 different EcoRl linkers (5' GGAATTCC, 5' CGGAATTCCG, and 5' CCGGAATTCCGG) was ligated onto each of the above pools, and the resulting ligation mixtures were digested with'EcoR1. The two pools were run on an agarose gel and fragments of apparent size of 800-4OOObp were isolated and purified on NA45 paper (Schliecher and Schuell) . At this point the preps were pooled, and 6ug of the pooled reactions was ligated with 5ug of CIP-treated Rl-cut pJG4-5 (Gyuris et al., submitted). Ligation products were purified and introduced into E coli "Sure13 cells (Stratagene, Inc.) K-12 (mcrA ~(mcrBC hs~MSmrr)177, endAl, supE44, thi-1, lambda-, gyrA96, relAl, lac, recB,-recJ, sbcC, umuC::Tn5_(kanR),uvrC/ F'[proAB, laclq, lacZ delta M15] : :Tnl0 (tetR) by electroporation.3-5 X 10e6 transformants were collected from 70 l5cm plates. The pooled cells were' then grown in LB for 5 hours, and plasmid DNA prepared from them by alkaline lysis followed by CsCl gradient purification.
Roger Brent 6 June 1993
617 726 5925
617 726 6893 (fax)
brent@opal .mgh.harvard. edu.
Serum Starved WI-38 library Made by Claude Sardet and Jeno Gyuris.
25 x 150mm dishes of recent passage (just purchased) WI-38 from the ATCC were grown to extreme confluence in DMEM + 15% 560 heat treated fetal calf serum and then shifted to DMEM + 0.1% FCS for three days. This should have given about 3 X lOexp7 cells/dish from which a total of 25mg total mRNA was extracted by Vanadium acid phenol treatment (Current Protocols in Molecular Biology). Total mRNA was divided into two aliquots, each of which was purified by two passages over 1 ml oligoDT columns (New England Biolabs) . This yielded 4Oug of poly A+ mRNA. 6ug of this mRNA was made into cDNA essentially as in the Gyuris Hela library except that, in the blunting step, Sarnet used RNAse A, RNAse H, T4 DNA polymerase, and E. coli DNA ligase, as described in CPMB. Blunt DNA had a NotI/EcoRl adaptor from Invitrogen (shown below) ligated ont9 its 5' end.
5' OH AATTCTGCGGCCGC
CACGCCGGCG 5? P04
Half of this cDNA was fractionated on Sepharose CL-'4b spin columns (Pharmacia), and Sardet picked the stuff that came through. This cDNA all looked bigger than about 6OObp and had an apparent median size of about 1.5 kb. The other half was run on a homemade spin column made from Sepharcryl 5-500 (Pharmacia) . All of this cDNA looked like it was bigger than l2OObp, and most of it looked to be about 3600bp. In both preps, the largest cDNAs'seemed to be about 8500bp. Both preps were combined, and all of this cDNA was ligated into a pJG4-5 backbone cut with Rl and XhoI. Ligation mix was introduced into storebought electroporation competent:DH1Ob(BRL), whose genotype is
F- mcrA del (mrr-hs~S-mcrBV) phi80 lacZdelM15 del lacX74 deoR recAl endA1 araDl39 del(ara,leu)7697 galU galK lambdarpsL nupG.
5.7 X lOexp6 transformants were collected on LB Amp plates. Of 30 library members examined, all but two had inserts, of "average" size of 1400 bp.'.' Among these 30, the two smallest inserts were, 300-500, and the biggest'was 3200. 2Omg of primary library DNA was made from an 8 liter prep grown in LB Amp.
Roger Brent
brent@opal .mgh harvard. edu
Interactor hunt-- protocol
The following procedure describes how to conduct an interactor hunt. Further information on yeast manipulation is available in (References) . The procedure described here utilizes information gained from hunts conducted in the lab by Jeno Gyuris, Russ Finley, Tony Zervos, Andy Mendelsohn, Ze'ev Paroush, Debbie Goff, Steve Hanes, Barak Cohen, Dimitri Krainc, Lakshmi Raj, Claude Sardet, Pierre Leopold, and Herbert Altman.
Interaction trap selection/screening protocol - Russ Finley
September, 1992
Updated 5/93
Yeast transformation
I currently favor using the protocol of Gietz et al., (1992, Nucleic
Acids Research vol. 20, pg 1425) which in our hands now gives
transformation efficiencies just under lOe5/ug DNA. To get this
performance, it is imperative that you use properly prepared carrier
DNA (sonicated, boiled salmon sperm DNA; see Schiestl and Gietz,
1989, Current Genetics -16:339-3A6)
Library DNA should be introduced into yeast that already harbor the URA3+ reporter and H153+ bait plasmids. Thus, yeast should be grown in ura-his- glucose CM dropout medium before transformation to maintain selection for these plasmids. I perform transformations in several individual eppendorf tubes (one for each 24cm X 24cm transformation plate) to reduce the likelihood of contaminating the entire transformation. It is important not to use excess transforming DNA per aliquot of competent yeast, otherwise each competent cell will take up several plasmids. We have found that adding DMSO to 10% (of the volume) just prior to adding the PEG solution increces efficiencies several fold. Also, it improves efficiency to resuspend the transformed yeast in sterile water rather than TE before plating Onto ura- his- trp- CM dropout glucose plates, or to plate PEG along with the yeast. It helps to make the plates with 25g agar/liter to make them extra hard. We use sterile glass beads. (e.g.Fisher #11-312b, 4mm diameter) to help spread the transformation mixture on the plates.
[Note: An alternative to the following protocol is to plate the transformation directly onto the selection plates (urahis-trp leu-, gal/raff) after incubation in gal/raff liquid for'a' few hours to induce the GAL promoter. We are currently comparing these two methods.]
To make it easier to determine the number of transformants, make dilutions (with water, not TE) from a few transformations and plate on normal sized ura- his- trp- CM dropout glucose plates. I strive for 200,000 transformants per 24cm X 24cm plate. This corresponds to lug per eppendorf tube if your transformation efficiency is 0.5 x lOe5 transformants per ug DNA.
Harvest trpnsformants
Incubate the plates for about 48 hours at 30 C, or until colonies
are about 1 mm in diameter. Cool plates at 4 C for a few hours
to harden agar. Use a sterile glass microscope slide (and sterile
technique) to scrape the yeast from the plate, trying not to scrape
any agar. The yeast can be collected from the glass slide by wiping
it on the lip of a sterile 50 ml Falcon tube. Wash the cells 2x
with TE. It is best to pellet the cells each time in a sterile
round bottom polypropylene tube at 2500 rpm for 4 mm. so they
may be easily resuspended The pellet volume for 500,000 transformants
will be about 8 ml. Resuspend the cells thoroughly in 1 pellet
volume of glycerol solution (Current Protocols in Molecular Biology:
65% glycerol (vol/vol), 0.1 M MgSO4, 25 mM Tris pH 8.0). Freeze
lml aliquots at -70 0C.
Determine plating efficiency
Remove an aliquot of transformed yeast and dilute 10-fold with
ura- his- trp- CM dropout gal/raff medium (2% galactose, 1% raffinose;
in this medium the yeast use raffinose as a carbon source but
the galactose still induces transcription from the GALl promoter)
. Incubate, shaking, at 30 oC for 4 hours to induce the GAL promoter.
Make serial dilutions using the above medium and plate on ura~his~trp-
gal/raff plates to determine the number of colony forming units
per aliquot of transformed yeast.
Plate onto selection plates
Induce the GAL promoter as~above~and plate < 10e6 colony forming
units per 100 mm plate ura- his- trp- leu- CM dropout gal/raff.
Pick Leu+ colonies and patch (or better yet, ~- -- -- streak for
single colonies) onto a new ura- his- trp- leu- CM dropout gal/raff
plate. The goal here is to purify the Leu+ yeast away from the
many leu- yeast on the original selection plate before putting
them on a master plate without leu selection. In order to test
for Gal inducibility of the Leu+ phenotype it is necessary to
first turn off the GAL promoter by growing on glucose master plates.
From the new ura- his- trp- leu- CM gal/raff plate patch to a
ura- his- trp- CM dropout glu master plate this will shut off
the GAL promoter. You can now replica from this glu plate to four
plates that will determine whether the Leu+ phenotype is galactose
dependent and whether it correlates with galactose-dependent betagalactosidase
activity. The four plates are 1. Ura- his- trp- CM dropout glu
X-GAL; 2. Ura- his- trp- CM dropout gal/raff xGAL; 3. Ura- his-
trp- leu- CM dropout glu; 4. Ura- his- trp- leu- CM dropout gal/raff.
It is important not to transfer too large a mass of yeast in the
replica.
Isolate library plasmid and test specificlty
Once you have isolated galactose-dependent LEU+ lacZ+ yeast, the
next step is to formally prove that the phenotype is due to the
library plasmid and to see that the interacting cDNAencoded protein
is specific for the bait. To do this we isolate the TRPl library
plasmid from the positive yeast and use it to retransform the
original bait strain, and other strains expressing unrelated baits
(for example, the LexA- bicoid derivative in pRFHMl)
If you have isolated a large number of positive yeast, it is useful to identify those that contain identical library plasmids so that the workload can be reduced. One quick way to determine which yeast have the same cDNA is to do a quick yeast miniprep, PCR the cDNA with primers derived from the vector pJG4-5, and cut the resulting PCR products with HaeIII and/or Alul. It is usually clear from this analysis which cDNAs are the same.We isolate the library plasmid by transforming E.coli strain KC8 (Gyruis et al, Cell, submitted, a gift of Kevin °Struhl) with the yeast miniprep DNA and selecting for complementation of the E.coli trpC mutation by yeast TRPl.
Russ Finley
finley@opal .mgh .harvard.edu
September 1992 updated May 1993
SEQUENCES
Primers for sequencing cDNAs in pJG4-5 from the 5' end can be derived from the sequence of the fusion moiety shown below ("5' sequence'1) For sequencing from the 3' end, primers can be derived from sequences in the ADHl (ADCl) terminator; the portion in pJG4-5 corresponds to sequences downstream of the natural HindIII site in the ADHl terminator. The following sequence ("3' sequence") is from this terminator region, about 40 bp downstream of the HindIll site. We have used primers complementary to this region for sequencing and PCR.
pJG4-5 3' sequence cDNA..XhoI..HindIII...ADHterm...5I...GTCTCCAATCAAGGTTGTCGGCTT
GTCTACCTTGCCAGAAATTTACGAAAAGATGGAAAAGGGTC 3'
pJG4-5 5' sequence
Fusion moiety from pJG4-5
(Linear) MAP of: N-Terminus.Txt check: 5219 from: 1 to: 3.45-
Acid activator.Txt Length: 345 June 14, 1991 13:26 Sequence of HinDIll-HinDIll fragment encoding the 5v40 nuclear localization sequence, the B42 activation region, and the HAl epitope from the influenza virus hemagglutinin gene, and containing the EcoRl and XhoI cloning sites.
With 177 enzymes: *
September 18, 1991 19:30
BEST TO FIND THIS YOURSELF! TRY GENBANK. ALSO GET THE SEQ FOR
pEG202
REQUESTING MATERIALS
If you would like plasmids and strains for the interaction trap you can e-mail
Russ Finley,
[email protected] harvard.edu, or send a fax to Roger Brent
617-726-6893
We have have now sent-out close to 400 shipments of interaction trap materials, and are currently sending off about 20 per week. To help us speed this up, please include the following information in your fax or e-mail:
Your name/ your lab.
Phone number
Fax number
email address
Mailing adaress including street address
Fedex account or other express shipping number (if you have one)
A description of what you are asking for. A description of the bait you intend to use
the name of the organism you wish to isolate interactors from (if you would like us to send you a library, which one?)
Russ Finley 20 September 1993
'1
REFERENCES
~Yeast techniques
1. Guthrie, C., and G.R. Fink 1991 Guide to yeast genetics and
molecular biology. Methods in Enzymology vol 194. Acedemic Press
Inc.
2. Ausubel, F.M., R. Brent, R.E. Kingston, D.D. Moore, J.G.
Seidmen, J.A. Smith, and K. Struhl. 1992 Current Protocols in
Molecular Biology, Greene Publishing Associates
Yeast transformation
Gietz, D., A. St. Jean, R.A. Woods, and R.H. Schiestl 1992
Nucleic Acids Research 20:1425
Bait return project
We ask you to send us your LexA fusion contructions for a project described below. Ideally, we would like to receive both the bait plasmid DNA and a derivative of yeast strain RFY206/pSH1834 into which you have introduced the bait plasmid. __
RFY206/pSH1834 is URA3+his3-. We enclose this strain with the standard kit (see requesting materials). You should maintain selection for the pSH1834 plasmid (a hypersensitive LexAop-lacZ reporter) by first steaking the strain onto Ura-glu plates.and then growing in Ura-glu liquid culture prior to transformation with your H1S3+ bait plasmid.
The more information about your LexA fusions that you can provide, the more useful it will be to us. If you can't manage the transformation we would be very grateful for the bait DNA.
You can send the strain and plasmid to
Ru~s Finley
Department of Molecular Bio~ogy
Massachusetts General Hospital
50 Blossom Street
Boston, Massachusetts 02114
USA
What this is about
We are isolating sets of proteins involved in cell cycle det~sions. Most of these are new, and we hope to use the interaction trap to gain a clue to the function of some of them, by identifying interactions between these proteins of unknown function and known bait proteins.
These baits do not need to be involved in cell cycle decisions: any interaction may be informative. If, for example we find that one of our new proteins interacts with known cytoskeletal ccmponents, or with a component of the replication apparatus, we may gain a clue to its function. Although this approach is of unproven value, we are quite excited about it, and we hope that if it works for our new cell cycle regulators it might be generalizable to newly identified proteins that may affect other cellular processes.
To this 'end, Russ Finley has devised simple plate mating conditions by which a putative interactor can be quickly tested against a very large number (lOOs or lOODs) of baits contained in ~