A paradigm shift

Document Sample
A paradigm shift Powered By Docstoc
					   Models of cellular regulation
• A genetic switch
• Lambda lysogeny/lysis
  – Three operator sites controlling two promoters P RM and
  – Cro and CI dimers bind to the operator sites, generating
    two antagonistic feedback loops
  – Cro dimer represses expression of CI, while CI
    represses Cro; bind to operators with different affinities
    and in opposite order
  – Concentration dependent logic
How do cells obtain signals from
• Uneven distribution of biomolecules among
• Stochastic gene expression has been
  observed in both eukaryotic and prokaryotic
• How do cells focus a signal for specific
  gene expression?
            A paradigm shift
• Reductionism  Integration
• System properties are determined by
  concentration of each component and
  reaction rates – even with steady state
  assumptions still a complex issue
• Model systems
  – Metabolism
  – Signal transduction
Genomics, proteomics, structural
       genomics, etc.
• Looking to reveal networks inherent to cell
  – Looking at models
  – Turning stochastic processes into deterministic
  Biological signaling occurs at
         multiple levels
• Intracellular signaling complexity results
  – Interactions between pathways
  – Compartmentalization
  – Signal channeling
• Many signaling components are membrane-
  bound, and there is a distinct dearth in our
  understanding of membrane biochemistry.
• Still, it has been readily identified that cells
  use compartments to derive specific
  microenvironments, which can offer distinct
  responses to the same signals
• Look at compartments as wires or
         Reaction channeling
• Central tenet of metabolism
• Compartments communicate via
• Consider transporters as switches (?)
  controlling the flow of signals down
• An intersection between cell biology and
   Fatty acids are activated and
transported into the mitochondria
Transduction by carnitine is the
major regulatory point of fatty
        acid oxidation
          Molecular scaffolds
• Once considered the function of rRNA
• Term used for a new class of signaling
  proteins that do not have information
  transfer capability of their own but interact
  with multiple signaling proteins in a
• “The scaffold provides an assembly line
  along which a series of enzymes process
  their substrates in a well-defined sequence
  and with an efficiency and specificity that
  are orders of magnitude higher than would
  be possible.”
  Approaches to the complexity
• Development of signaling databases (ie.
• Systematic cataloging of proteins, lipids,
  sugars, and other signaling molecules
  together with genomic data of model
     An example in modeling –
       metabolic phenomics
• “It is now clear that we need to develop
  creative approaches and technologies to use
  all of this information [genomics and
  proteomics] to explore and determine
  genome function. We must essentially take
  on the view of a gene that we began with
  over 50 years ago, wherein the focus was on
  the functional attributes of a gene within the
  context of the whole organism.”
• Even when multiple knockouts are
  generated, a surprising number of mutants
  result in no effect on growth.
• Flexibility in metabolic genotype –
  rerouting of metabolites
• Clear example given by PK knockout in E.
  Yet, some metabolic modeling
   and engineering successes
• Prediction and correlation of defined growth
• Glucose transporter confers heterotrophic
  growth upon a photosynthetic algae
• Check out PLAS
          Integrated circuits
• How do metabolic pathways communicate?
• How do signal transduction pathways illicit
  appropriate responses?
• Etc.
     Start with a simple model
• Michaelis-Menten
• Modeling interactions between adenosine
  receptor with adenylate cyclase with first
  order kinetics – Handout
       b-adrenergic receptors
• Integral membrane protein with 7 TM
  regions – serpentine receptor
• Epinephrine (or adrenaline) binds and
  causes a conformational change that
  stimulates a G protein, which in turn
  stimulates adenylyl cyclase
Epinephrine transduction
G Protein has a built-in timing
The adenylyl cyclase reaction
        Modeling this reaction
b-receptor is physically separated and
  activates the enzyme by “collision
• Modeled as a first order reaction in the
  presence of non-hydrolyzable GTP
• Expressing the results mathematically
  Activation of adenylate cyclase
           by adenosine
• In contrast to collision coupling, the adenosine
  receptor is modeled as permanently coupled to
  adenylate cyclase
• This predicts a distinct rate constant dependence
  for cyclase activity (cyclase activation)
• Adenosine activation of adenylate cyclase is
  predicted to be independent of receptor
  concentration (k3 is unaltered), but the maximum
  catalytic units will decrease upon receptor
           Braun and Levitzki
• Examine figure 3; o-adenosine is a competitive
  inhibitor that does not affect the catalytic rate
  regarding adenosine activation of adenylate
• This result is consistent with their model
• Additional support comes from independence of
  adenosine activation from membrane fluidity
• Relax, “permanent” means k3>>k1
    Use simple models to build
       complicated ones …
         Signal transduction
• Bhalla and Iyengar
• Signaling pathways are wires, since not
  separated by insulators – signaling
  molecules are distinct
A role for cAMP
Desensitization from persistent signal
      Other second messengers
• Phospholipase C cleaves membrane lipid
  phosphatidylinositol 4,5 bisphospate into
  two messengers diacylgllycerol and inositol
  1,4,5 trisphosphate (IP3)
• IP3 in turn activates release of calcium ions
  that act as a messenger and activate protein
  kinase C (numerous isozymes with tissue
  specific roles, for instance in cell division)
   PLC mediated
signal transduction
Regulation of cell cycle by
     protein kinases
Cyclin-dependent protein kinases
        control cell cycle
• By phosphorylating specific proteins at
  precise time intervals these kinases
  orchestrate the metabolic activities of the
  cell for cell division
• Heterodimers – one regulatory subunit
  (cyclin) and one catalytic subunit (cyclin-
  dependent protein kinase [CDK])_
Post-translational regulation through
  phosphorylation and proteolysis
Four mechanisms to control CDK
• Phosphorylation
   – Phosphorylate tyrosine prevents ATP binding
   – Removal of phosphate from tyrosine and
     phosphorylation of threonine allows substrate binding
• Controlled degradation
   – Feedback loop involving DBRP
• Regulated synthesis of CDKs and cyclins
   – MAPK mediated activation of Jun and Fos
• Inhibition of CDK
   – Specific proteins such as p21 bind and inactivate CDK
Observe variations in the activities
of specific CDKs during cell cycle
Whither MAPK?
       MAPK kinase cascade
• Many signals stimulate MAPK kinase
  cascade, but the wire is well conserved in
  biology – Handout
• Why does MAPK kinase use three kinases
  instead of one?
• Allows conversion of graded inputs into
  switch-like outputs
Regulation of passage from G1 to S
 Neuron function and
  signal transduction
• Voltage- and ligand-gated
• ion channels
Allosteric effectors of protein
     Glutamate receptor
          Forming memories
• Mini-review handout
           Integrating circuits
• Circuits exhibit synergy within a cellular context
• Bhalla and Iyengar modeling signal transduction
  in the brain and long-term potentiation (LTP) (Fig
• PKC activates MAPK, while MAPK helps
  activate PKC (Figure 8.16)
 Why does it take 100 minutes of
   5 nM EGF to reach LTP?
• 10 min at 5 nM or 100 min at 2 nM EGF is
  insufficient for LTP (Fig 8.18)
• Fig 8.19 result of determining concentration
  dependence of MAPK activation of PKC and the
• Three intersection points – MM 8.2 “Cobweb”
  – A indicates high activity for both enzymes
  – B indicates low activity for both
  – T is threshold stimulation, if EGF is sufficient to
    activate either PKC or MAPK above T – both will
    reach A (T serves as a switch between A and B)
           Turning off LTP
• Use a phosphatase to knock MAPK below
• AA (arachidonic acid) generated by PLA2
  persists, which makes it hard to turn off
• Takes awhile to de-phosphatase
       Integrating more circuits
• Start with MAPK circuit
• Add calcium activation, etc.
• Result in Figure 8.23
  –   PKC
  –   MAPK
  –   cAMP
  –   Calcium
          A network algorithm
• Derived in analogous fashion to protein interaction
• Use RegulonDB as training set
• Set up a matrix where the score = 1 if an operon
  (j) encodes a transcription factor that regulates
  another operon (I) to detect network motifs
• Random model – maintain number of connections
  but partners are chosen randomly
       Applied to several model
•   Biochemistry
•   Ecology
•   Neurobiology
•   Engineering (WWW)
     The similarity of networks
• Although components are unique among
  these models, the topologic properties of
  various networks share similarities.
• Universal organizing principles apply to all
  networks from cell to WWW?
      Does gene order matter?
• Cis-regulatory elements, proteins, and
  messengers are integrated into biological
• Does gene location in the genome affect the
• Genome evolution – gene order does matter
  that’s why we observe synteny
             Gene order in T7
• T7 produces 59 proteins from 56 genes…only 33
  have known function
• T7 infection is unique, first 850 bp are inserted,
  transcription begins, then the remainder is pulled
• E. coli polymerase pulls the first 15% of genomic
  DNA into the cell at ~45 bp/sec through
  transcription at 5 promoters– what a cool
  molecular machine
                 Gene 1
• T7 RNA polymerase
• Uses 17 different promoters in the
  remaining 85% of genome
• Pulls at a rate of 200 bp per second.
• What happens if Gene 1 is moved elsewhere
  on the genome?
           In silico analysis
• Measured optimal time for phage-induced
  lysis for 72 distinct T7 genomes
• Some genotypes were better than others
• T7 is suboptimal? Where’s the data?
• Three phage genome constructs were
  generated and tested at positions 1.7, 3.8
  and 12 (controls had random DNA inserted
  at these positions or a late promoter inserted
  early in genome)
• Little agreement between predicted and
  experimental data
              Systems biology
• Watson School of Biological Sciences at CSH
• “…The systems approach defines all of th
  eelements in a system and then studies how each
  behaves in relation to the others as the system is
  functioning. Ultimately the systems approach
  requires mathematical model which will both
  describe the nature of the system and its systems
  “Systems Biology Superstars”
• Integration of multiple -omes:
  – Metabolomics
  – Proteomics
  – Genomics
• “Looking at individual silos of genomics,
  proteomics, or metabolomics is akin to
  using a laser pointer in a dark office to
  describe its contents…”
Galactose metabolism in yeast as
          an example
• Define all genes in the genome and the subset of
  genes, proteins and other molecules constituting
  the galactose a model
• Perturb each pathway component using genetics or
  environmental challenges
• Utilize microarrays and ICAT to collect gene
  expression data
• Refine model
        Functional genomics
• Grew wild type and deletion strains and
  assessed gene expression via microarrays
• Used Northerns as controls
• How reliable are microarrays?
• Measured 289 proteins using ICAT, only 30
  observed differences; 15 of which showed
  no change in RNA levels, post-
  transcriptional control
            Going system…
• Ideker uses Fields protein interaction data to
  identify 997 mRNA and 15 proteins whose
  expression is altered by galactose
• Discovery questions 7-9 in Chapter 9
• I relent on the writing: “Typically, if good
  data conflict with your model, trust your
Clinical Proteomics
       Identifying Biomarkers
• Recall a web videocast regarding this topic
  from NIH
• A test using mass spectroscopic analysis of
  proteins predicts ovarian cancer 95% of the
  time – is this good? 20 out of 100,000
  women afflicted – Bayes Rule
         Personalized medicine
• Herceptin – aimed for 25-30% of women with
  breast cancer
• Drug development opens door for diagnostics, not
  vice-versa (diagnostics not being generated for
  diagnostics sake)
• Isn’t there some legislation looking to pass that
  relieves pharmaceutical companies from legal
  responsibilities for their products and side effects?
        Investigating Disease
• Clinical Presentation
  – Biopsy and Labs
• Family pedigree
  – DNA is inherited
• Karyotyping and Linkage analysis
• DNA sequence analysis
            Duchenne’s MD
• Following this process – identified
  dystrophin as causative gene/protein
• How do you work towards a solution?
  – An animal model – mice
  – Dystrophin has a paralog – utrophin, which is
    ubiquitously expressed, distinct domains within
    these proteins lead to distinct localization and
    protein interactions
    Finding dystrophin’s molecular
• Immunoprecipitation – leads to Figure 10.7
• But then what?
• (graph theory and critical nodes in Figure 10.8)
• Mutations in any of the genes encoding these gene
  products can lead to MD
• Ensuing discussion on the inaccuracy of one gene-
  one function-one phenotype posed as attending a
              Drug Delivery
• Viral vehicles
  – Provides specificity for cell type, can be
    performed in vitro or in vivo
  – Liposomes offer an alternative
• Protein carriers
  – Protein-transduction domain
• Nucleic acids
            Drug dilemmas
• The inefficacy of aspirin and Cox proteins
• Want to inhibit Cox-2, which produces
  prostaglandins that result in PAIN
• However, aspirin has 100X more affinity
  for Cox-1 than Cox-2

Shared By: