Transcript BioSci 145B Lecture #10 6/7/2005

BioSci 145B Lecture #10 6/7/2005
• Bruce Blumberg
– 2113E McGaugh Hall - office hours Wed 10-11 AM (or by appointment)
– phone 824-8573
– [email protected]
• TA – Suman Verma [email protected]
– 2113 McGaugh Hall, 924-6873, 3116
BioSci 145B lecture 10
page 1
©copyright
Bruce Blumberg 2004. All rights reserved
Library-based methods to map protein-protein interactions (contd)
• Phage display screening (a.k.a. panning)
– requires a library that expresses
inserts as fusion proteins with a
phage capsid protein
• most are M13 based
• some lambda phages used
– prepare target protein
• as affinity matrix
• or as radiolabeled probe
– test for interaction with library members
• if using affinity matrix you purify phages from a mixture
• if labeling protein one plates fusion protein library and probes with
the protein
– called receptor panning based on similarity with panning for
gold
BioSci 145B lecture 10
page 2
©copyright
Bruce Blumberg 2004. All rights reserved
Library-based methods to map protein-protein interactions (contd)
• Phage display screening (a.k.a. panning) (contd)
– advantages
• stringency can be manipulated
• if the affinity matrix approach works the cloning could go rapidly
– disadvantages
• Fusion proteins bias the screen against full-length cDNAs
• Multiple attempts required to optimize binding
• Limited targets possible
• may not work for heterodimers
• unlikely to work for complexes
• panning can take many months for each screen
– Greg Weiss in Chemistry is local expert
BioSci 145B lecture 10
page 3
©copyright
Bruce Blumberg 2004. All rights reserved
Mapping protein-protein interactions (contd)
• Two hybrid screening
– originally used in yeast, now
other systems possible
– prepare bait - target protein
fused to DBD (GAL4) usual
• stable cell line is commonly
used
– prepare fusion protein library
with an activation domain - prey
– Key factor required for success is
no activation domain in bait!
– approach
• transfect library into cells and either
select for survival or activation of
reporter gene
• purify and characterize positive clones
BioSci 145B lecture 10
page 4
©copyright
Bruce Blumberg 2004. All rights reserved
Mapping protein-protein interactions (contd)
• Two-hybrid screening (contd)
– Can be easily converted to
genome wide searching by
making haploid strains, each
containing one candidate
interactor
– Mate these and check for
growth or expression of reporter
gene
Bait plasmid
Prey plasmid
If interact, reporter expressed
and/or
Yeast survive
BioSci 145B lecture 10
page 5
©copyright
Bruce Blumberg 2004. All rights reserved
Molecular Interaction Screening - A New Approach to Protein Function
• Principle
– small pools of cDNAs are transcribed and translated in vitro to produce
pools of proteins that may be assayed in a variety of ways
• EMSA, co-ip, FRET, SPA
– cDNAs identified by protein function
• Starting material arrayed in 384-well plates
– Robotically pool source plates into daughter 96/384-well plates
• Pool size is optimizable - 96 works well
• Grow bacteria, prepare DNA,
TNT -> labeled protein
• Perform functional assay (SPA)
• Unpool positive wells into
components and rescreen
– Positive pools have known
composition
• only one second level
screen is required
BioSci 145B lecture 10
page 6
©copyright
Bruce Blumberg 2004. All rights reserved
Automated Molecular Interaction Screening
• Why do it this way?
– arbitrary size and complexity of target is possible
– Normalized cDNA pool -> representation of rare messages
– numerous possible endpoint assays
• radioactive, fluorescent, luminescent
– saturation screening of genome is feasible
– two screening steps to pure cDNA of interest in ~2 weeks
BioSci 145B lecture 10
page 7
©copyright
Bruce Blumberg 2004. All rights reserved
Large scale mapping of protein-protein interactions
• GST (glutathione-S-transferase)
pulldown assay
– Or other purification wherein one
protein is tagged and complex of
proteins binding to it is recovered
– Purify complexes from cells
– Characterize complexes by massspectrometry
– Iteratively build up a map of
protein interactions from such
complexes
BioSci 145B lecture 10
page 8
©copyright
Bruce Blumberg 2004. All rights reserved
Genomics - linking biological variation to disease pathophysiology
Biological system
Tissues
Populations
Cells
Animal strains
Patients
Clinical trial volunteers
Tissues
Stimulated / non-stimulated
Resistant / susceptible
Cases / controls
Responders / non-responders
Normal / treated-diseased
Multivariate!
Experimental system
protein
DNA
Variant between
individuals / populations
Genome sequence
Genotyping variation
RNA
Variant between tissues
Variant between tissues
Differential display
cDNA sequence (EST)
DNA microarrays
2D-electrophoresis / LC
Mass spectroscopy
( Yeast 2 hybrid )
What are genomic approaches to aid in these studies?
BioSci 145B lecture 10
page 9
©copyright
Bruce Blumberg 2004. All rights reserved
The rise of -omics
• The -omics revolution of science
– http://www.genomicglossaries.com/content/omes.asp
• What does it all mean?
– Transcriptomics – large scale gene profiling (usually microarray)
– Proteomics – study of complement of expressed proteins
– Functional genomics – very vague term, typically encompasses many
others
– Structural genomics – prediction of structure and interactions from
sequence
– Pharmacogenomics – transcriptional profiling of response to drug
treatment – often looking for genetic basis of differences
– Toxicogenomics – transcriptional profiling of response to toxicants (often
includes pharmacogenomics
• Seeks mechanistic understanding of toxic response
– Metabolomics – analysis of total metabolite pool ("metabolome") to
reveal novel aspects of cellular metabolism and global regulation
– Interactomics – genome wide study of macromolecular interactions,
physical and genetic are included.
BioSci 145B lecture 10
page 10
©copyright
Bruce Blumberg 2004. All rights reserved
The rise of –omics (contd)
• What do we want to know for drug development?
– How do individuals respond to drugs differently – pharmacogenomics
– How do individuals respond differently to toxicants - toxicogenomics
Target identification
Protein
Assay
Target validation
All of them!!
Compound library
Hit identification (HTS)
Hit
Genes
Hit to lead (Lead identification)
Lead optimization
Candidate drug
Effort
Clinical trials
BioSci 145B lecture 10
page 11
©copyright
Bruce Blumberg 2004. All rights reserved
Toxicogenomics
• Lump pharmacogenomics and toxicogenomics together in the context of drug
development
• Toxicology is the study of effects of toxicant exposure
– Traditional toxicology focuses on exposure, dose, effect
– “dose makes the poison” – overly simplistic and probably incorrect
• Mechanistic Toxicology (academic and regulatory)
– Investigative toxicology
• Hypothesis generation for grants and studies
– Risk assessment
• Understanding the mechanism of toxicity at the molecular level
• EPA and NIEHS very concerned with this
• Predictive toxicology
– Compound avoidance
• Elimination of liabilities (pharma, chemical industry)
– Compound selection
• Select compound with least toxic liability from a series (pharma)
– Compound management
• Tailor conventional studies and perform timely investigational
toxicology studies
BioSci 145B lecture 10
page 12
©copyright
Bruce Blumberg 2004. All rights reserved
Toxicogenomics (contd)
• Where predictive and mechanistic toxicology fit into drug development
– The road from hit to marketed drug is long
– 8/9 drug candidates fail due to toxic effects or unfavorable profile of
metabolism
Drug
Discovery
PreClinical
Testing
Clinical
Development
FDA
Mechanistic studies
Pattern-based
Mechanism-based
Predictive screens
BioSci 145B lecture 10
page 13
©copyright
Bruce Blumberg 2004. All rights reserved
Phase
IV
Toxicogenomics (contd)
• Bioinformatics ties together toxicogenomic studies
• Overall goal is predictive, personalized medicine
– Provide personalized prescriptions to best help each patient
• Especially cancer therapy
Infrastructure
Clinical and experimental material
SNP Genotyping
Genome data
DNA
Novel targets
Novel pathways
Novel diagnostic indicators
Mining
Novel biomarkers
Predictive toxicology
Modelling Predictive pharmacology
Analysis
Microarray data
EST / cDNA data
RNA
protein
Proteomics
Predictive medicine
function
BioSci 145B lecture 10
Functional readouts
Metabolic space
Chemistry space
page 14
©copyright
Bruce Blumberg 2004. All rights reserved
Novelty, mechanism & prediction - toxicogenomics
Can we replace
animal studies with
genomics analyses?
Rat tissues
Normal and treated
Timecourses
BioSci 145B lecture 10
page 15
©copyright
Bruce Blumberg 2004. All rights reserved
Toxicogenomics (contd)
• What is toxicogenomics good for?
– Obtaining a high level view of a biological system
– Rapid generation of response profiles to
• Unravel mechanisms
• Discriminate among compounds
– Signature of exposures?
– Probably not a single method to identify toxicity
• Problems that must be solved
– Interlab variation – different labs use slightly different methods and get
results that may not be strictly applicable
• Japanese solution is to designate a single lab for entire country
– Most genes change expression at high doses of exposure
• Relevant?
BioSci 145B lecture 10
page 16
©copyright
Bruce Blumberg 2004. All rights reserved
Genomic technology - implications
• Genetics and reverse genetics
– gene transfer and selection technology speeds up genetic analysis by
orders of magnitude
– virtually all conceivable experiments are now possible
• all questions are askable
• BUT should all questions be asked?
– much more straightforward to understand gene function using knockouts
and transgenics
• gene sequences are coming at an unprecedented rate from the
genome projects
• Knockouts and transgenics remain very expensive to practice
– other yet undiscovered technologies will be required to
understand gene function.
BioSci 145B lecture 10
page 17
©copyright
Bruce Blumberg 2004. All rights reserved
Genomic technology – implications (contd)
• Clinical genetics
– Molecular diagnostics are becoming very widespread as genes are
matched with diseases
• huge growth area for the future
• big pharma is dumping billions into diagnostics
– room for great benefit and widespread abuse
• diagnostics will enable early identification and treatment of diseases
• but insurance companies will want access to these data to maximize
profits
BioSci 145B lecture 10
page 18
©copyright
Bruce Blumberg 2004. All rights reserved
Genomic technology – implications (contd)
• gene therapy
– new viral vector technology is making this a reality
• efficient transfer and reasonable regulation possible
– long lag time from laboratory to clinic, still working with old technology
in many cases
– The Biotech Death of Jesse Gelsinger. Sheryl Gay Stolberg, NY Times,
Sunday Magazine, 28 Nov 99
• http://www.frenchanderson.org/history/biotech.pdf
• protein engineering
– not as widely appreciated as more glamorous techniques such as gene
therapy and transgenic crops
– better drugs, e.g., more stable insulin, TPA for heart attacks and
strokes, etc.
– more efficient enzymes (e.g. subtilisin in detergents)
– safe and effective vaccines
• just produce antigenic proteins rather than using inactivated or
attenuated organisms to reduce undesirable side effects
BioSci 145B lecture 10
page 19
©copyright
Bruce Blumberg 2004. All rights reserved
Genomic technology – implications (contd)
• metabolite engineering
– enhanced microbial synthesis of valuable products
• eg indigo (jeans)
• vitamin C
– generation of entirely new small molecules
• transfer of antibiotic producing genes to related species yields new
antibiotics (badly needed)
– reduction of undesirable side reactions
• faster more efficient production of beer
• plants as producers of specialty chemicals
– underutilized because plant technology lags behind techniques in animals
• But regulations are strict (Monsanto)
– plants as factories to produce materials more cheaply and efficiently
• especially replacements for petrochemicals
– plants and herbs are the original source of many pharmaceutical products
• engineer them to overproduce desirable substances
BioSci 145B lecture 10
page 20
©copyright
Bruce Blumberg 2004. All rights reserved
Genomic technology – implications (contd)
• transgenic food
– gene transfer techniques have allowed the creation of desirable
mutations into animals and crops of commercial value
• disease resistance (various viruses)
• pest resistance (Bt cotton)
• Pesticide, herbicide and fungicide resistance
• growth hormone and milk production
– effective but necessary?
– negative implications – “Frankenfoods”
• pesticide and herbicide resistance lead to much higher use of toxic
compounds
• results are not predictable due to small datasets
• at least one herbicide (bromoxynil) for which resistance was
engineered has since been banned
• Atrazine is becoming highly controversial
BioSci 145B lecture 10
page 21
©copyright
Bruce Blumberg 2004. All rights reserved
2004 Final Examination
1. (10 points) Remember our old friend Mars burroughsii from the midterm? The
pathogenic eukaryotic microorganism that escaped into Newport Beach and is now
spreading along the Pacific Coast? Recall also that we have generated a full genomic
sequence and a full set of ESTs. We have learned that a protein on the M. burroughsii
surface binds to the human CadF protein in order to gain entry into the cell and that
polymorphisms in the CadF gene confer a degree of resistance to M. burroughsii
infection in a subset of the population. It is now time to figure out just how M.
burroughsii exerts its pathogenic effects in the cell. A postdoc in your lab hypothesizes
that additional M. burroughsii proteins interact with human cellular proteins to cause
pathogenesis. Comparisons of different M. burroughsii isolates have identified 50
potential genes that are involved in pathogenecity.
a) (5 points) How would you identify the full complement of human proteins in
intestinal cells that the 50 M. burroughsii proteins interact with?
b) (5 points) Assume that one of the human proteins identified in a) is quite
different between humans and mice. Recall that M. burroughsii seems only to
infect humans and not mice. How could you go about creating a mouse model for
studying M. burroughsii pathogenesis? Assume that M. burroughsii can infect
mouse cells by binding to CadF just fine, but that they are apparently not
pathogenic due to this single protein difference between mice and humans.
BioSci 145B lecture 10
page 22
©copyright
Bruce Blumberg 2004. All rights reserved
a) Possible answers are a proteomic approach or large-scale two hybrid. For the
proteomic approach, it would be necessary to tag the 50 proteins so that
they could be identified, put these back into the cells, purify complexes and
identify the components. For the two-hybrid approach, one would need to
make suitable bait from each of the 50 proteins in one yeast mating type
strain and then cross these with prey from another strain and look for
interactions.
b) You would want to use targeted mutagenesis to create a mouse that
expresses the human protein in place of the corresponding mouse protein.
Alternatively, you could simply make a transgenic mouse expressing the
human protein but this might not work if there is interference between the
human and mouse versions of the protein.
BioSci 145B lecture 10
page 23
©copyright
Bruce Blumberg 2004. All rights reserved
2. (5 points) OK, you now have a mouse model for M. burroughsii pathogenesis
that can be used to study the molecular mechanisms involved and 50
candidate M. burroughsii proteins that are involved in pathogenecity. How
would you go about systematically disrupting these genes one by one to
determine where each fit into the pathogenesis cascade?
The best approach would be to create M. burroughsii that are deficient in
the proteins using homologous recombination. Alternatively, you could try
something like siRNA expressed under the control of a ubiquitously active
promoter to knock down the proteins and test the effects of these proteins.
A proteomic approach wherein you mutate the proteins and test the effect
on the complexes identified in question 1a could also be employed.
BioSci 145B lecture 10
page 24
©copyright
Bruce Blumberg 2004. All rights reserved
3. (5 points) Great, now you have identified which of the M. burroughsii
proteins act where to induce a pathogenic response in infected cells. These
proteins will be good candidate drug targets. Several companies have
identified candidate drugs for testing by screening compound libraries
against these targets but first wish to determine the profiles of gene
expression in response to these drugs in mice and humans. How would you go
about determining these response profiles and how they differ in hepatocytes
from different human donors and wild type mice vs. your mouse model from
1b above?
The best approach would be to treat human hepatocytes from different
donors with the candidate drugs and evaluate the profile of gene expression
using microarray analysis. The profiles could then be compared with each
other to develop a sense of how humans vary in response to the drugs and
whether the profile of gene expression corresponds to known toxic profiles
obtained from toxicogenomic studies. A similar approach using wild-type and
your model mouse would give the desired results.
BioSci 145B lecture 10
page 25
©copyright
Bruce Blumberg 2004. All rights reserved
4. (10 points) You are a molecular biologist who really loves eating shrimp after
work as much as you enjoy working in the lab. To combine your two loves,
you have decided to create a biotech company based on producing the
largest shrimp anywhere. You took an invertebrate biology course at UCI and
learned that shrimp and related animals express an unidentified growth
inhibitory gene when cultured at high density. In answering the questions
below, assume that all of the standard sorts of methods we discussed in class
will work in shrimp (except targeted gene disruption). Assume that shrimp
heterozygous for disruption of the growth inhibitory gene will be large
enough to be detected but not big enough for your company.
a) (4 points) Describe an approach that will allow you to create mutations in
the growth inhibitory gene
b) (4 points) Describe how you will clone and/or identify the gene that has
been mutated?
c) (2 points) How will you figure out where the gene is expressed to begin to
understand how it works?
BioSci 145B lecture 10
page 26
©copyright
Bruce Blumberg 2004. All rights reserved
a) you don’t have any genome sequence available from this organism therefore
siRNA or targeted gene disruption will not work. You could try chemical
mutagenesis but that will not put you in a position to answer b and c. The
best way is to use gene trapping to identify a line that has enhanced growth
abilities.
b) if you created a gene trap line in a) then all you have to do is clone the
region surrounding the insertion and sequence it to identify the disrupted
gene
c) once you have established the gene trap line, the expression patterns of the
reporter gene should tell you where it is expressed. Alternatively, you could
use in situ hybridization but this might not be sensitive enough to detect
expression everywhere.
BioSci 145B lecture 10
page 27
©copyright
Bruce Blumberg 2004. All rights reserved
5. (5 points) In answering question 4 above, you have identified a chunk of
genomic sequence that appears to correspond to the gene responsible for
growth inhibition. You identify a large BAC clone containing 150 kb of
sequence and sequence it completely.
a) (3 points) Describe how you will identify where the introns, exons,
splicing signals, and the putative promoter are in this gene.
b) (2 points) How will you determine how many mRNAs are expressed from
this region of genomic DNA?
a) A bioinformatics approach to identify introns, exons, splicing signals,
etc might be good enough. You could also use the genomic sequence to
isolate cDNAs that could then be compared with the genomic sequence
to identify the introns, exons and putative promoter
b) you could use the genomic DNA to hybridize to northern blots or
microarrays or a bioinformatics approach to identify predicted ORFs
BioSci 145B lecture 10
page 28
©copyright
Bruce Blumberg 2004. All rights reserved
6. (5 points) The jumbo jumbo shrimp that your company is selling are bringing
in serious $ in profits. The CEO decides that improving the product even
further would be a wise investment and commissions an EST project to be
performed in your laboratory. 15,000 different transcripts are identified and
glass slide microarrays are prepared. To more fully characterize the crowding
response, your lab is contracted to identify all of the genes that are
upregulated and downregulated by crowding in aquaculture. Describe
generally how you will accomplish this.
One good approach would be to use the glass slide microarrays to
characterize upregulated and downregulated transcripts. You would prepare
mRNA from shrimp that are grown in a crowded environment and from a
non-crowded environment. These mRNAS would be used to prepare
fluorescently labeled probes that would be hybridized to the microarrays.
Analysis of the microarray data would identify candidate upregulated and
downregulated genes that could be tested by QRT-PCR to distinguish which
ones are truly regulated by crowding and which are not.
BioSci 145B lecture 10
page 29
©copyright
Bruce Blumberg 2004. All rights reserved