Transformer Proteins against Cancer
Larry H. Bernstein, MD, FCAP, Curator
LPBI
Revision 1/13/20165
GEN News Highlights: “Transformer” Proteins with Role in Cancer
http://www.genengnews.com/gen-news-highlights/transformer-proteins-with-role-in-cancer/81252133/
Investigators at the Scripps Research Institute (TSRI) and St. Jude Children’s Research Hospital say they have found how a protein involved in cancer twists and morphs into different structures.
“We’re studying basic biophysics, but we believe the complexity and rules we uncover for the physics of protein disorder and folding could one day also be used for better designs of therapeutics,” said Ashok Deniz, Ph.D., associate professor at TSRI.
The study (“Asymmetric Modulation of Protein Order-Disorder Transitions by Phosphorylation and Partner Binding”), published in Angewandte Chemie, focuses on the nucleophosmin (NPM1) protein, which has many functions and, when mutated, has been shown to interfere with cells’ normal tumor suppressing ability. NPM1 has been implicated in cancers such as non-Hodgkin lymphoma and acute myelogenous leukemia.
Previous research led by study collaborators Richard Kriwacki, Ph.D., and Diana Mitrea, Ph.D., at St. Jude had demonstrated that a section of NPM1, called the N-terminal domain (Npm-N), doesn’t have a defined, folded structure. Instead, the protein morphs between two forms: a one-subunit disordered monomer and a five-subunit folded pentamer.
Until now, the mechanism behind this transformation was unknown, but scientists believed this monomer-pentamer equilibrium could be important for the protein’s location and functioning in the cell. To shed light on how this transformation occurred, Dr. Deniz and his colleagues used a combination of three techniques—single-molecule biophysics, fluorescence resonance energy transfer (FRET), and circular dichroism, which enabled them to study individual molecules and collections of molecules. Single-molecule methods are especially useful for such studies because they can uncover important information that remains hidden in conventional studies.
The researchers found that the transformation can proceed through more than one pathway. In one pathway, the transformation begins when the cell sends signals to attach phosphoryl groups to NPM1. Such phosphorylation prompts the ordered pentamer to become disordered and likely causes NPM1 to shuttle outside the cell’s nucleus. A meeting with a binding partner can mediate the reverse transformation to a pentamer.
When NPM1 does become a pentamer again under these conditions, which likely causes it to move back to the nucleolus, it takes a different path instead of just retracing its earlier steps.
Priya Banerjee, Ph.D., an American Heart Association-supported postdoctoral research associate at TSRI and the first author of the study, compared these complicated transitions to the morphing of a “Transformers” toy, where a robot can become a car and then a jet. “Phosphorylation and partner-binding are like different cellular switches driving these changes,” said Dr. Banerjee.
According to Dr. Banerjee, the new study also reveals many intermediate states between monomer and pentamer structures and that these states can be manipulated or “tuned” by changing conditions such as salt levels, phosphorylation, and partner binding, which may explain how cells regulate the protein’s multiple functions. The researchers said future studies could shed more light on the biological functions of these different structures and how they might be used in future cancer therapies.
The team added that combining the three techniques used in this study, plus a novel protein-labeling technique for single-molecule fluorescence, could be a useful strategy for studying other unstructured, “intrinsically disordered proteins” (IDPs), which are involved in a host of cellular functions, as well as neurodegenerative disease, heart disease, infectious disease, type 2 diabetes and other conditions.
Single ‘Transformer’ Proteins
http://www.technologynetworks.com/Proteomics/news.aspx?ID=186996
“We’re studying basic biophysics, but we believe the complexity and rules we uncover for the physics of protein disorder and folding could one day also be used for better designs of therapeutics,” said TSRI Associate Professor Ashok Deniz, senior author of the new study along with Richard Kriwacki, faculty member at St. Jude.
The study focuses on a protein called nucleophosmin (NPM1). This protein has many functions and, when mutated, has been shown to interfere with cells’ normal tumor-suppressing ability. NPM1 has been implicated in cancers such as non-Hodgkin lymphoma and acute myelogenous leukemia.
Previous research led by study collaborators Kriwacki and Diana Mitrea at St. Jude had shown that a section of NPM1, called the N-terminal domain (Npm-N), doesn’t have a defined, folded structure. Instead, the protein morphs between two forms: a one-subunit disordered monomer and a five-subunit folded pentamer.
Until now, the mechanism behind this transformation was unknown, but scientists believed this monomer-pentamer equilibrium could be important for the protein’s location and functioning in the cell.
To shed light on how this transformation occurred, Deniz and his colleagues used an innovative combination of three techniques—single-molecule biophysics, fluorescence resonance energy transfer (FRET) and circular dichroism—which enabled them to study individual molecules and collections of molecules. Single-molecule methods are especially useful for such studies because they can uncover important information that remains hidden in conventional studies.
Remarkably, the researchers found that the transformation can proceed through more than one pathway. In one pathway, the transformation begins when the cell sends signals to attach phosphoryl groups to NPM1. This modification, called phosphorylation, prompts the ordered pentamer to become disordered and likely causes NPM1 to shuttle outside the cell’s nucleus. A meeting with a binding partner can mediate the reverse transformation to a pentamer.
Interestingly, when NPM1 does become a pentamer again under these conditions, which likely causes it to move back to the nucleolus, it takes a different path instead of just retracing its earlier steps.
Priya Banerjee, an American Heart Association-supported postdoctoral research associate at TSRI and the first author of the study, compared these complicated transitions to the morphing of a “Transformers” toy, where a robot can become a car and then a jet. “Phosphorylation and partner-binding are like different cellular switches driving these changes,” said Banerjee.
Banerjee said the new study also reveals many intermediate states between monomer and pentamer structures—and that these states can be manipulated or “tuned” by changing conditions such as salt levels, phosphorylation and partner binding, which may explain how cells regulate the protein’s multiple functions. The researchers said future studies could shed more light on the biological functions of these different structures and how they might be used in future cancer therapies.
The researchers added that combining the three techniques used in this study, plus a novel protein-labeling technique for single-molecule fluorescence, could be a useful strategy for studying other unstructured, “intrinsically disordered proteins” (IDPS). IDPS are involved in a host of cellular functions, as well as neurodegenerative disease, heart disease, infectious disease, type 2 diabetes and other conditions.
TSRI and St. Jude Scientists Study Physics of Single ‘Transformer’ Proteins with Role in Cancer
http://www.scripps.edu/news/press/2015/20151221deniz.html
December 21, 2015 – A new study led by scientists at The Scripps Research Institute (TSRI) and St. Jude Children’s Research Hospital shows how a protein involved in cancer twists and morphs into different structures.
“We’re studying basic biophysics, but we believe the complexity and rules we uncover for the physics of protein disorder and folding could one day also be used for better designs of therapeutics,” said TSRI Associate Professor Ashok Deniz, senior author of the new study along with Richard Kriwacki, faculty member at St. Jude.
The study, published recently in the journal Angewandte Chemie, focuses on a protein called nucleophosmin (NPM1). This protein has many functions and, when mutated, has been shown to interfere with cells’ normal tumor suppressing ability. NPM1 has been implicated in cancers such as non-Hodgkin lymphoma and acute myelogenous leukemia.
Previous research led by study collaborators Kriwacki and Diana Mitrea at St. Jude had shown that a section of NPM1, called the N-terminal domain (Npm-N), doesn’t have a defined, folded structure. Instead, the protein morphs between two forms: a one-subunit disordered monomer and a five-subunit folded pentamer.
Until now, the mechanism behind this transformation was unknown, but scientists believed this monomer-pentamer equilibrium could be important for the protein’s location and functioning in the cell.
To shed light on how this transformation occurred, Deniz and his colleagues used an innovative combination of three techniques—single-molecule biophysics, fluorescence resonance energy transfer (FRET) and circular dichroism—which enabled them to study individual molecules and collections of molecules. Single-molecule methods are especially useful for such studies because they can uncover important information that remains hidden in conventional studies.
Remarkably, the researchers found that the transformation can proceed through more than one pathway. In one pathway, the transformation begins when the cell sends signals to attach phosphoryl groups to NPM1. This modification, called phosphorylation, prompts the ordered pentamer to become disordered and likely causes NPM1 to shuttle outside the cell’s nucleus. A meeting with a binding partner can mediate the reverse transformation to a pentamer.
Interestingly, when NPM1 does become a pentamer again under these conditions, which likely causes it to move back to the nucleolus, it takes a different path instead of just retracing its earlier steps.
Asymmetric Modulation of Protein Order-Disorder Transitions by Phosphorylation and Partner Binding
Priya R. Banerjee, Diana M. Mitrea, Richard W. Kriwacki, Ashok, A. Deniz
http://dx.doi.org:/10.1002/anie.201507728
As for many intrinsically disordered proteins, order–disorder transitions in the N-terminal oligomerization domain of the multifunctional nucleolar protein nucleophosmin (Npm-N) are central to its function, with phosphorylation and partner binding acting as regulatory switches. However, the mechanism of this transition and its regulation remain poorly understood. In this study, single-molecule and ensemble experiments revealed pathways with alternative sequences of folding and assembly steps for Npm-N. Pathways could be switched by altering the ionic strength. Phosphorylation resulted in pathway-specific effects, and decoupled folding and assembly steps to facilitate disorder. Conversely, binding to a physiological partner locked Npm-N in ordered pentamers and counteracted the effects of phosphorylation. The mechanistic plasticity found in the Npm-N order–disorder transition enabled a complex interplay of phosphorylation and partner-binding steps to modulate its folding landscape.
Conditionally and Transiently Disordered Proteins: Awakening Cryptic Disorder To Regulate Protein Function
Related articles:
Select item 264270603.
Design, Synthesis and Evaluation of 2,5-Diketopiperazines as Inhibitors of the MDM2-p53 Interaction.
Pettersson M, Quant M, Min J, Iconaru L, Kriwacki RW, Waddell MB, Guy RK, Luthman K, Grøtli M.
PLoS One. 2015 Oct 1;10(10):e0137867. doi: 10.1371/journal.pone.0137867. eCollection 2015.
A phosphorylation switch controls the spatiotemporal activation of Rho GTPases in directional cell migration
Xuan Cao, Tomonori Kaneko, Jenny S. Li, An-Dong Liu, et al.
Nature Communications 2015; 6(7721) http://dx.doi.org:/10.1038/ncomms8721
Although cell migration plays a central role in development and disease, the underlying molecular mechanism is not fully understood. Here we report that a phosphorylation-mediated molecular switch comprising deleted in liver cancer 1 (DLC1), tensin-3 (TNS3), phosphatase and tensin homologue (PTEN) and phosphoinositide-3-kinase (PI3K) controls the spatiotemporal activation of the small GTPases, Rac1 and RhoA, thereby initiating directional cell migration induced by growth factors. On epidermal growth factor (EGF) or platelet-derived growth factor (PDGF) stimulation, TNS3 and PTEN are phosphorylated at specific Thr residues, which trigger the rearrangement of the TNS3–DLC1 and PTEN–PI3K complexes into the TNS3–PI3K and PTEN–DLC1 complexes. Subsequently, the TNS3–PI3K complex translocates to the leading edge of a migrating cell to promote Rac1 activation, whereas PTEN–DLC1 translocates to the posterior for localized RhoA activation. Our work identifies a core signalling mechanism by which an external motility stimulus is coupled to the spatiotemporal activation of Rac1 and RhoA to drive directional cell migration.
Protein Phosphorylation is an Important Tool… Chapter 13. in Protein Phosphorylation in Human Health
Roberta Fraschini, Erica Raspelli and Corinne Cassani
http://cdn.intechopen.com/pdfs-wm/38809.pdf
Protein phosphorylation is a reversible posttranslational modification that can modulate protein role in several physiological processes in almost every possible way. These include modification of its intrinsic biological activity, subcellular location, half-life and binding with other proteins. Protein phosphorylation is particularly important for the regulation of key proteins involved in the control of cell cycle progression.
Protein phosphorylation is the covalent binding of a phosphate group to some critical residues of the polypeptide. The phosphorylation state of a protein is given by a balance between the activity of protein kinases and protein phosphatases. Eukaryotic protein kinases transfer phosphate groups (PO43-) from ATP to an hydroxyl group of the lateral chain of specific serine, threonine or tyrosine residues on peptide substrates. In simple eukaryotic cells, like yeasts, Ser/Thr kinases are more common, while more complex eukaryotic cells, like human cells, have many Tyr kinases. Protein kinases recognize their substrates specifically and their active site consists of an activation loop and a catalytic loop between which substrates bind. Protein kinases differ from each other in the structure of their catalitic domain. A second class of enzymes, protein phosphatases, is responsible for the reverse reaction in which phosphate groups are removed from a protein. Phosphatases gain specificity by binding protein cofactors which facilitate binding to specific phosphoproteins. The active phosphatase often consists of a complex of the phosphatase catalytic subunit and a regulatory subunit.
The phosphorylation of specific residues induces structural changes that regulate protein functions by modulating protein folding, substrate affinity, stability and activity. For example, phosphorylation can cause switch-like changes in protein function, which can also lead to major modifications in the catalytic function of enzymes, including kinases and phosphatases. In addition, protein phosphorylation often leads to rearrangement in the structure of the protein that can induce changes in interacting partners or subcellular localization.
Phosphorylation acts as a molecular switch for many regulatory events in signalling pathways that drive cell division, proliferation, differentiation and apoptosis. In order to ensure an appropriate balance of protein phosphorylation, the cell can compartmentalize both protein kinases and phosphatases. Another kind of regulation can be achieved by the spatial distribution of kinases and phosphatases, that creates a gradient of phosphorylated substrates across different subcellular compartments. This spatial separation can also control the activity of other proteins or enzymes and the occurrence of other posttranslational modifications.
Molecular Biology of the Cell. 4th edition. Protein Function. http://www.ncbi.nlm.nih.gov/books/NBK26911/
We have seen that each type of protein consists of a precise sequence of amino acids that allows it to fold up into a particular three-dimensional shape, or conformation. But proteins are not rigid lumps of material. They can have precisely engineered moving parts whose mechanical actions are coupled to chemical events. It is this coupling of chemistry and movement that gives proteins the extraordinary capabilities that underlie the dynamic processes in living cells.
In this section, we explain how proteins bind to other selected molecules and how their activity depends on such binding. We show that the ability to bind to other molecules enables proteins to act as catalysts, signal receptors, switches, motors, or tiny pumps. The examples we discuss in this chapter by no means exhaust the vast functional repertoire of proteins. However, the specialized functions of many of the proteins you will encounter elsewhere in this book are based on similar principles.
All Proteins Bind to Other Molecules
The biological properties of a protein molecule depend on its physical interaction with other molecules. Thus, antibodies attach to viruses or bacteria to mark them for destruction, the enzyme hexokinase binds glucose and ATP so as to catalyze a reaction between them, actin molecules bind to each other to assemble into actin filaments, and so on. Indeed, all proteins stick, or bind, to other molecules. In some cases, this binding is very tight; in others, it is weak and short-lived. But the binding always shows great specificity, in the sense that each protein molecule can usually bind just one or a few molecules out of the many thousands of different types it encounters. The substance that is bound by the protein—no matter whether it is an ion, a small molecule, or a macromolecule— is referred to as a ligand for that protein (from the Latin word ligare, meaning “to bind”).
The ability of a protein to bind selectively and with high affinity to a ligand depends on the formation of a set of weak, noncovalent bonds—hydrogen bonds, ionic bonds, and van der Waals attractions—plus favorable hydrophobic interactions (see Panel 2-3, pp. 114–115). Because each individual bond is weak, an effective binding interaction requires that many weak bonds be formed simultaneously. This is possible only if the surface contours of the ligandmolecule fit very closely to the protein, matching it like a hand in a glove ().
The selective binding of a protein to another molecule. Many weak bonds are needed to enable a protein to bind tightly to a second molecule, which is called aligand for the protein. A ligand must therefore fit precisely into a protein’s binding (more…)
The region of a protein that associates with a ligand, known as the ligand’s binding site, usually consists of a cavity in the protein surface formed by a particular arrangement of amino acids. These amino acids can belong to different portions of the polypeptide chain that are brought together when the protein folds (). Separate regions of the protein surface generally provide binding sites for different ligands, allowing the protein’s activity to be regulated, as we shall see later. And other parts of the protein can serve as a handle to place the protein in a particular location in the cell—an example is the SH2 domain discussed previously, which is often used to move a protein containing it to sites in theplasma membrane in response to particular signals.
The binding site of a protein. (A) The folding of the polypeptide chain typically creates a crevice or cavity on the protein surface. This crevice contains a set of amino acid side chains disposed in such a way that they can make noncovalent bonds only (more…)
Although the atoms buried in the interior of the protein have no direct contact with the ligand, they provide an essential scaffold that gives the surface its contours and chemical properties. Even small changes to the amino acids in the interior of a protein molecule can change its three-dimensional shape enough to destroy a binding site on the surface.
The Details of a Protein’s Conformation Determine Its Chemistry
Proteins have impressive chemical capabilities because the neighboring chemical groups on their surface often interact in ways that enhance the chemical reactivity of amino acid side chains. These interactions fall into two main categories.
First, neighboring parts of the polypeptide chain may interact in a way that restricts the access of water molecules to aligand binding site. Because water molecules tend to form hydrogen bonds, they can compete with ligands for sites on the protein surface. The tightness of hydrogen bonds (and ionic interactions) between proteins and their ligands is therefore greatly increased if water molecules are excluded. Initially, it is hard to imagine a mechanism that would exclude a molecule as small as water from a protein surface without affecting the access of the ligand itself. Because of the strong tendency of water molecules to form water–water hydrogen bonds, however, water molecules exist in a large hydrogen-bonded network (see Panel 2-2, pp. 112–113). In effect, a ligand binding site can be kept dry because it is energetically unfavorable for individual water molecules to break away from this network, as they must do to reach into a crevice on a protein’s surface.
Second, the clustering of neighboring polar amino acid side chains can alter their reactivity. If a number of negatively charged side chains are forced together against their mutual repulsion by the way the protein folds, for example, the affinity of the site for a positively charged ion is greatly increased. In addition, when amino acid side chains interact with one another through hydrogen bonds, normally unreactive side groups (such as the –CH2OH on the serine shown in) can become reactive, enabling them to enter into reactions that make or break selected covalent bonds.
An unusually reactive amino acid at the active site of an enzyme. This example is the “catalytic triad” found in chymotrypsin, elastase, and other serine proteases (see Figure 3-14). The aspartic acid side chain (Asp 102) induces the histidine (more…)
The surface of each protein molecule therefore has a unique chemical reactivity that depends not only on which amino acid side chains are exposed, but also on their exact orientation relative to one another. For this reason, even two slightly different conformations of the same protein molecule may differ greatly in their chemistry.
Sequence Comparisons Between Protein Family Members Highlight Crucial Ligand Binding Sites
As we have described previously, many of the domains in proteins can be grouped into families that show clear evidence of their evolution from a common ancestor, and genome sequences reveal large numbers of proteins that contain one or more common domains. The three-dimensional structures of the members of the same domain family are remarkably similar. For example, even when the amino acid sequence identity falls to 25%, the backbone atoms in a domain have been found to follow a common protein fold within 0.2 nanometers (2 Å).
These facts allow a method called “evolutionary tracing” to be used to identify those sites in a protein domain that are the most crucial to the domain’s function. For this purpose, those amino acids that are unchanged, or nearly unchanged, in all of the known protein family members are mapped onto a structural model of the three-dimensional structure of one family member. When this is done, the most invariant positions often form one or more clusters on the protein surface, as illustrated in for the SH2 domain described previously (see Panel 3-2, pp. 138–139). These clusters generally correspond to ligand binding sites.
The evolutionary trace method applied to the SH2 domain. (A) Front and back views of a space-filling model of the SH2 domain, with evolutionarily conserved amino acids on the protein surface colored yellow, and those more toward the protein interior colored (more…)
The SH2 domain is a module that functions in protein–protein interactions. It binds the protein containing it to a second protein that contains a phosphorylated tyrosine side chain in a specific amino acid sequence context, as shown in . The amino acids located at the binding site for the phosphorylated polypeptide have been the slowest to change during the long evolutionary process that produced the large SH2 family of peptide recognition domains. Becausemutation is a random process, this result is attributed to the preferential elimination during evolution of all organisms whose SH2 domains became altered in a way that inactivated the SH2-binding site, thereby destroying the function of the SH2 domain.
In this era of extensive genome sequencing, many new protein families have been discovered whose functions are unknown. By identifying the critical binding sites on a three-dimensional structure determined for one family member, the above method of evolutionary tracing is being used to help determine the functions of such proteins.
Proteins Bind to Other Proteins Through Several Types of Interfaces
Proteins can bind to other proteins in at least three ways. In many cases, a portion of the surface of one protein contacts an extended loop of polypeptide chain (a “string”) on a second protein (). Such a surface–string interaction, for example, allows the SH2 domain to recognize a phosphorylated polypeptide as a loop on a second protein, as just described, and it also enables a protein kinase to recognize the proteins that it will phosphorylate (see below).
Three ways in which two proteins can bind to each other. Only the interacting parts of the two proteins are shown. (A) A rigid surface on one protein can bind to an extended loop of polypeptide chain (a “string”) on a second protein. (B)(more…)
A second type of protein–protein interface is formed when two α helices, one from each protein, pair together to form acoiled-coil (). This type of protein interface is found in several families of gene regulatory proteins, as discussed in Chapter 7.
The most common way for proteins to interact, however, is by the precise matching of one rigid surface with that of another (). Such interactions can be very tight, since a large number of weak bonds can form between two surfaces that match well. For the same reason, such surface–surface interactions can be extremely specific, enabling aprotein to select just one partner from the many thousands of different proteins found in a cell.
Molecular Motors
Perhaps the most fascinating proteins that associate with the cytoskeleton are the molecular motors called motor proteins. These remarkable proteins bind to a polarized cytoskeletal filament and use the energy derived from repeated cycles of ATP hydrolysis to move steadily along it. Dozens of different motor proteins coexist in every eucaryotic cell. They differ in the type of filament they bind to (either actin or microtubules), the direction in which they move along the filament, and the “cargo” they carry. Many motor proteins carry membrane-enclosed organelles—such as mitochondria, Golgi stacks, or secretory vesicles—to their appropriate locations in the cell. Other motor proteins cause cytoskeletal filaments to slide against each other, generating the force that drives such phenomena as muscle contraction, ciliary beating, and cell division.
Cytoskeletal motor proteins that move unidirectionally along an oriented polymer track are reminiscent of some other proteins and protein complexes discussed elsewhere in this book, such as DNA and RNA polymerases, helicases, and ribosomes. All of these have the ability to use chemical energy to propel themselves along a linear track, with the direction of sliding dependent on the structural polarity of the track. All of them generate motion by coupling nucleosidetriphosphate hydrolysis to a large-scale conformational change in a protein, as explained in Chapter 3.
The cytoskeletal motor proteins associate with their filament tracks through a “head” region, or motor domain, that binds and hydrolyzes ATP. Coordinated with their cycle of nucleotide hydrolysis and conformational change, the proteins cycle between states in which they are bound strongly to their filament tracks and states in which they are unbound. Through a mechanochemical cycle of filament binding, conformational change, filament release, conformational relaxation, and filament rebinding, the motor protein and its associated cargo move one step at a time along the filament (typically a distance of a few nanometers). The identity of the track and the direction of movement along it are determined by the motor domain (head), while the identity of the cargo (and therefore the biological function of the individual motor protein) is determined by the tail of the motor protein.
In this section, we begin by describing the three groups of cytoskeletal motor proteins. We then describe how they work to transport membrane-enclosed organelles or to change the shape of structures built from cytoskeletal filaments. We end by describing their action in muscle contraction and in powering the whiplike motion of structures formed from microtubules.
Actin-based Motor Proteins Are Members of the Myosin Superfamily
The first motor protein identified was skeletal muscle myosin, which is responsible for generating the force for muscle contraction. This myosin, called myosin II (see below) is an elongated protein that is formed from two heavy chains and two copies of each of two light chains. Each of the heavy chains has a globular head domain at its N-terminus that contains the force-generating machinery, followed by a very long amino acid sequence that forms an extended coiled-coil that mediates heavy chain dimerization (). The two light chains bind close to the N-terminal head domain, while the long coiled-coil tail bundles itself with the tails of other myosin molecules. These tail-tail interactions result in the formation of large bipolar “thick filaments” that have several hundred myosin heads, oriented in opposite directions at the two ends of the thick filament ().
Myosin II. (A) A myosin II molecule is composed of two heavy chains (each about 2000 amino acids long (green) and four light chains (blue). The light chains are of two distinct types, and one copy of each type is present on each myosin head. Dimerization (more…)
The myosin II bipolar thick filament. (A) Electron micrograph of a myosin II thick filament isolated from frog muscle. Note the central bare zone, which is free of head domains. (B) Schematic diagram, not drawn to scale. The myosin II molecules aggregate (more…)
Each myosin head binds and hydrolyses ATP, using the energy of ATP hydrolysis to walk toward the plus end of anactin filament. The opposing orientation of the heads in the thick filament makes the filament efficient at sliding pairs of oppositely oriented actin filaments past each other. In skeletal muscle, in which carefully arranged actin filaments are aligned in “thin filament” arrays surrounding the myosin thick filaments, the ATP-driven sliding of actin filaments results in muscle contraction (discussed later). Cardiac and smooth muscle contain myosins that are similarly arranged, although they are encoded by different genes.
When a muscle myosin is digested by chymotrypsin and papain, the head domain is released as an intact fragment (called S1). The S1 fragment alone can generate filament sliding in vitro, proving that the motor activity is contained completely within the head ().
Direct evidence for the motor activity of the myosin head. In this experiment, purified S1 myosin heads were attached to a glass slide, and then actin filaments labeled with fluorescent phalloidin were added and allowed to bind to the myosin heads. (A) (more…)
It was initially thought that myosin was present only in muscle, but in the 1970’s, researchers found that a similar two-headed myosin protein was also present in nonmuscle cells, including protozoan cells. At about the same time, other researchers found a myosin in the freshwater amoeba Acanthamoeba castellanii that was unconventional in having a motor domain similar to the head of muscle myosin but a completely different tail. This molecule seemed to function as a monomer and was named myosin I (for one-headed); the conventional myosin was renamed myosin II (for two-headed).
Subsequently, many other myosin types were discovered. The heavy chains generally start with a recognizable myosin motor domain at the N-terminus, and then diverge widely with a variety of C-terminal tail domains (). The new types of myosins include a number of one-headed and two-headed varieties that are approximately equally related to myosin I and myosin II, and the nomenclature now reflects their approximate order of discovery (myosin III through at least myosin XVIII). The myosin tails (and the tails of motor proteins generally) have apparently diversified during evolution to permit the proteins to dimerize with other subunits and to interact with different cargoes.
Myosin superfamily tree. (A) A family tree for a few of the many known members of the myosin superfamily. The length of the lines separating individual family members indicates the amount of difference in the amino acid sequence of the motor domain. Groups (more…)
Some myosins (such as VIII and XI) have been found only in plants, and some have been found only in vertebrates (IX). Most, however, are found in all eucaryotes, suggesting that myosins arose early in eucaryotic evolution. The yeastSaccharomyces cerevisiae contains five myosins: two myosin Is, one myosin II, and two myosin Vs. One can speculate that these three types of myosins are necessary for a eucaryotic cell to survive and that other myosins perform more specialized functions in multicellular organisms. The nematode C. elegans, for example, has at least 15 myosin genes, representing at least seven structural classes; the human genome includes about 40 myosin genes.
All of the myosins except one move toward the plus end of an actin filament, although they do so at different speeds. The exception is myosin VI, which moves toward the minus end.
The exact functions for most of the myosins remain to be determined. Myosin II is always associated with contractile activity in muscle and nonmuscle cells. It is also generally required for cytokinesis, the pinching apart of a dividing cell into two daughters (discussed in Chapter 18), as well as for the forward translocation of the body of a cell during cell migration. The myosin I proteins contain a second actin–binding site or a membrane-binding site in their tails, and they are generally involved in intracellular organization and the protrusion of actin-rich structures at the cell surface. Myosin V is involved in vesicle and organelle transport. Myosin VII is found in the inner ear in vertebrates, and certain mutations in the gene coding for myosin VII cause deafness in mice and humans.
There Are Two Types of Microtubule Motor Proteins: Kinesins and Dyneins
Kinesin is a motor protein that moves along microtubules. It was first identified in the giant axon of the squid, where it carries membrane-enclosed organelles away from the neuronal cell body toward the axon terminal by walking toward the plus end of microtubules. Kinesin is similar structurally to myosin II in having two heavy chains and two light chains per active motor, two globular head motor domains, and an elongated coiled-coil responsible for heavy chain dimerization. Like myosin, kinesin is a member of a large protein superfamily, for which the motor domain is the only common element (). The yeastSaccharomyces cerevisiae has six distinct kinesins. The nematode C. elegans has 16 kinesins, and humans have about 40.
Kinesin and kinesin-related proteins. (A) Structures of four kinesin superfamily members. As in the myosin superfamily, only the motor domains are conserved. Conventional kinesin has the motor domain at the N-terminus of the heavy chain. The middle domain (more…)
There are at least ten families of kinesin-related proteins, or KRPs, in the kinesin superfamily. Most of them have the motor domain at the N-terminus of the heavy chain and walk toward the plus end of the microtubule. A particularly interesting family has the motor domain at the C-terminus and walks in the opposite direction, toward the minus end of the microtubule. Some KRP heavy chains lack a coiled-coil sequence and seem to function as monomers, analogous to myosin I. Some others are homodimers, and yet others are heterodimers. At least one KRP (BimC) can self-associate through the tail domain, forming a bipolar motor that slides oppositely oriented microtubules past one another, much as a myosin II thick filament does for actin filaments. Most kinesins carry a binding site in the tail for either a membrane-enclosed organelle or another microtubule. Many of the kinesin superfamily members have specific roles in mitotic and meiotic spindle formation and chromosome separation during cell division.
The dyneins are a family of minus-end-directed microtubule motors, but they are unrelated to the kinesin superfamily. They are composed of two or three heavy chains (that include the motor domain) and a large and variable number of associated light chains. The dynein family has two major branches (). The most ancient branch contains thecytoplasmic dyneins, which are typically heavy-chain homodimers, with two large motor domains as heads. Cytoplasmic dyneins are probably found in all eucaryotic cells, and they are important for vesicle trafficking, as well as for localization of the Golgi apparatus near the center of the cell. Axonemal dyneins, the other large branch, include heterodimers and heterotrimers, with two or three motor-domain heads, respectively. They are highly specialized for the rapid and efficient sliding movements of microtubules that drive the beating of cilia and flagella (discussed later). A third, minor, branch shares greater sequence similarity with cytoplasmic than with axonemal dyneins but seems to be involved in the beating of cilia.
Dyneins. Freeze-etch electron micrographs of a molecule of cytoplasmic dynein and a molecule of ciliary (axonemal) dynein. Like myosin II and kinesin, cytoplasmic dynein is a two-headed molecule. The ciliary dynein shown has three heads. Note that the (more…)
Dyneins are the largest of the known molecular motors, and they are also among the fastest: axonemal dyneins can move microtubules in a test tube at the remarkable rate of 14 μm/sec. In comparison, the fastest kinesins can move their microtubules at about 2–3 μm/sec.
The Structural Similarity of Myosin and Kinesin Indicates a Common Evolutionary Origin
The motor domain of myosins is substantially larger than that of kinesins, about 850 amino acids compared with about 350. The two classes of motor proteins track along different filaments and have different kinetic properties, and they have no identifiable amino acid sequence similarities. However, determination of the three-dimensional structure of the motor domains of myosin and kinesin has revealed that these two motor domains are built around nearly identical cores (). The central force-generating element that the two types of motor proteins have in common includes the site of ATP binding and the machinery necessary to translate ATP hydrolysis into an allosteric conformational change. The differences in domain size and in the choice of track can be attributed to large loops extending outward from this central core. These loops include the actin-binding and microtubule-binding sites, respectively.
X-ray crystal structures of myosin and kinesin heads. The central nucleotide-binding domains of myosin and kinesin (shaded in yellow) are structurally very similar. The very different sizes and functions of the two motors are due to major differences (more…)
An important clue to how the central core is involved in force generation has come from the observation that the motor core also bears some structural resemblance to the nucleotidebinding site of the small GTPases of the Ras superfamily. As discussed in Chapter 3 (see Figure 3-74), these proteins exhibit distinct conformations in their GTP-bound (active) and GDP-bound (inactive) forms: mobile “switch” loops in the nucleotide-binding site are in close contact with the γ-phosphate in the GTP-bound state, but these loops swing out when the hydrolyzed γ-phosphate is released. Although the details of the movement are different for the two motor proteins, and ATP rather than GTP is hydrolyzed, the relatively small structural change in the active site—the presence or absence of a terminal phosphate—is similarly amplified to cause a rotation of a different part of the protein. In kinesin and myosin, a switch loop interacts extensively with those regions of the protein involved in microtubule and actin binding, respectively, allowing the structural transitions caused by the ATP hydrolysis cycle to be relayed to the polymer-binding interface. The relay of structural changes between the polymer-binding site and the nucleotide hydrolysis site seems to work in both directions, since theATPase activity of motor proteins is strongly activated by binding to their filament tracks.
Motor Proteins Generate Force by Coupling ATP Hydrolysis to Conformational Changes
Although the cytoskeletal motor proteins and GTP-binding proteins both use structural changes in their nucleoside-triphosphate-binding sites to produce cyclic interactions with a partner protein, the motor proteins have a further requirement: each cycle of binding and release must propel them forward in a single direction along a filament to a newbinding site on the filament. For such unidirectional motion, a motor protein must use the energy derived from ATP binding and hydrolysis to force a large movement in part of the protein molecule. For myosin, each step of the movement along actin is generated by the swinging of an 8.5-nm-long α helix, or lever arm (see ), which is structurally stabilized by the binding of light chains. At the base of this lever arm next to the head, there is a piston-like helix that connects movements at the ATP-binding cleft in the head to small rotations of the so-called converter domain. A small change at this point can swing the helix like a long lever, causing the far end of the helix to move by about 5.0 nm. These changes in the conformation of the myosin are coupled to changes in its binding affinity for actin, allowing the myosin head to release its grip on the actin filament at one point and snatch hold of it again at another. The full mechanochemical cycle of nucleotide binding, nucleotide hydrolysis, and phosphate release (which causes the “power stroke”) produces a single step of movement (). In the myosin VI subfamily of myosins, which move backward (toward the minus end of the actin filament), the converter domain probably lies in a different orientation, so that the same piston-like movement of the small helix causes the lever arm to rotate in the opposite direction.
The cycle of structural changes used by myosin to walk along an actin filament. (Based on I. Rayment et al., Science 261:50–58, 1993. © AAAS.)
In kinesin, instead of the rocking of a lever arm, the small movements of switch loops at the nucleotide–binding siteregulate the docking and undocking of the motor head domain to a long linker region that connects this motor head at one end to the coiled-coil dimerization domain at the other end. When the front (leading) kinesin head is bound to amicrotubule before the power stroke, its linker region is relatively unstructured. On the binding of ATP to this bound head, its linker region docks along the side of the head, which throws the second head forward to a position where it will be able to bind a new attachment site on the protofilament, 8 nm closer to the microtubule plus end than the binding site for the first head. The nucleotide hydrolysis cycles in the two heads are closely coordinated, so that this cycle of linker docking and undocking can allow the two-headed motor to move in a hand-over-hand (or head-over-head) stepwise manner ().
Comparison of the mechanochemical cycles of kinesin and myosin II. The shading in the two circles representing the hydrolysis cycle indicates the proportion of the cycle spent in attached and detached states for each motor protein. (A) Summary of the (more…)
The coiled-coildomain seems both to coordinate the mechanochemical cycles of the two heads (motor domains) of thekinesin dimer and to determine its directionality of movement. Recall that whereas most members of the kinesin superfamily, with their motor domains at the N-terminus, move toward the plus end of the microtubule, a few superfamily members have their motor domains at the C-terminus and move toward the minus end. Since the motor domains of these two types of kinesins are essentially identical, how can they move in opposite directions? The answer seems to lie in the way in which the heads are connected. In high-resolution images of forward-walking and backward-walking members of the kinesin superfamily bound to microtubules, the heads that are attached to the microtubule are essentially indistinguishable, but the second, unattached heads are oriented very differently. This difference in tilt apparently biases the next binding site for the second head, and thereby determines the directionality of motor movement ().
Orientation of forward- and backward-walking kinesin superfamily proteins bound to microtubules. These images were generated by fitting the structures of the free motor-protein dimers (determined by x-ray crystallography) onto a lower resolution image (more…)
Although both myosin and kinesin undergo analogous mechanochemical cycles, the exact nature of the coupling between the mechanical and chemical cycles is different in the two cases (see ). For example, myosin without any nucleotide is tightly bound to its actin track, in a so-called “rigor” state, and it is released from this track by the association of ATP. In contrast, kinesin forms a rigor-like tight association with a microtubule when ATP is bound to the kinesin, and it is hydrolysis of ATP that promotes release of the motor from its track.
Thus, cytoskeletal motor proteins work in a manner highly analogous to GTP-binding proteins, except that in motor proteins the small protein conformational changes (a few tenths of a nanometer) associated with nucleotide hydrolysis are amplified by special protein domains—the lever arm in the case of myosin and the linker in the case of kinesin—to generate large-scale (several nanometers) conformational changes that move the motor proteins stepwise along their filament tracks. The analogy between the GTPases and the cytoskeletal motor proteins has recently been extended by the observation that one of the GTP-binding proteins—the bacterial elongation factorG—translates the chemical energy of GTP hydrolysis into directional movement of the mRNAmolecule on the ribosome.
Motor Protein Kinetics Are Adapted to Cell Functions
The motor proteins in the myosin and kinesin superfamilies exhibit a remarkable diversity of motile properties, well beyond their choice of different polymer tracks. Most strikingly, a single dimer of conventional kinesin moves in a highly processive fashion, traveling for hundreds of ATPase cycles along a microtubule without dissociating. Skeletal muscle myosin II, in contrast, cannot move processively and makes just one or a few steps along an actin filament before letting go. These differences are critical for the motors’ various biological roles. A small number of kinesin molecules must be able to transport a mitochondrion all the way down a nerve cellaxon, and therefore require a high level of processivity. Skeletal muscle myosin, in contrast, never operates as a single molecule but rather as part of a huge array of myosin II molecules. Here processivity would actually inhibit biological function, since efficient muscle contraction requires that each myosin head perform its power stroke and then quickly get out of the way, to avoid interfering with the actions of the other heads attached to the same actin filament.
There are two reasons for the high degree of processivity of kinesin movement. The first is that the mechanochemical cycles of the two motor heads in a kinesin dimer are coordinated with each other, so that one kinesin head does not let go until the other is poised to bind. This coordination allows the motor protein to operate in a hand-over-hand fashion, never allowing the organelle cargo to diffuse away from the microtubule track. There is no apparent coordination between the myosin heads in a myosin II dimer. The second reason for the high processivity of kinesin movement is that kinesin spends a relatively large fraction of its ATPase cycle tightly bound to the microtubule. For both kinesin and myosin, the conformational change that produces the force-generating working stroke must occur while the motor protein is tightly bound to its polymer, and the recovery stroke in preparation for the next step must occur while the motor is unbound. But as we have seen in , myosin spends only about 5% of its ATPase cycle in the tightly bound state and is unbound the rest of the time.
What myosin loses in processivity it gains in speed; in an array in which many motor heads are interacting with the sameactin filament, a set of linked myosins can move their filament a total distance equivalent to 20 steps during a single cycle time, while kinesins can move only two. Thus, myosins can typically drive filament sliding much more rapidly than kinesins, even though they hydrolyze ATP at comparable rates and take molecular steps of comparable length.
Within each motor protein class, movement speeds vary widely, from about 0.2 to 60 μm/sec for myosins, and from about 0.02 to 2 μm/sec for kinesins. These differences arise from a fine-tuning of the mechanochemical cycle. The number of steps that an individual motor molecule can take in a given time, and thereby the velocity, can be increased by either increasing the motor protein’s intrinsic ATPase rate or decreasing the proportion of cycle time spent bound to the filament track. Moreover, the size of each step can be changed by either changing the length of the lever arm (for example, the lever arm of myosin V is about three times longer than the lever arm of myosin II) or the angle through which the helix swings (). Each of these parameters varies slightly among different members of the myosin and kinesin families, corresponding to slightly different protein sequences and structures. It is assumed that the behavior of each motor protein, whose function is determined by the identity of the cargo attached through its tail-domain, has been fine-tuned during evolution for speed and processivity according to the specific needs of the cell.
The effect of lever arm length on the step size for a motor protein. The lever arm of myosin II is much shorter than the lever arm of myosin V. The power stroke in the head swings their lever arms through the same angle, so myosin V is able to take a (more…)
Motor Proteins Mediate the Intracellular Transport of Membrane-enclosed Organelles
A major function of cytoskeletal motors in interphase cells is the transport and positioning of membrane-enclosed organelles. Kinesin was originally identified as the protein responsible for fast axonal transport, the rapid movement of mitochondria, secretory vesicle precursors, and various synapse components down the microtubule highways of theaxon to the distant nerve terminals. Although organelles in most cells need not cover such long distances, their polarized transport is equally necessary. A typical microtubule array in an interphase cell is oriented with the minus ends near the center of the cell at the centrosome, and the plus ends extending to the cell periphery. Thus, centripetal movements of organelles toward the cell center require the action of minus-end-directed motor proteins such as cytoplasmic dynein, whereas centrifugal movements toward the periphery require plus-end-directed motors such as kinesins.
The role of microtubules and microtubule motors in the behavior of intracellular membranes is best exemplified by the part they play in organizing the endoplasmic reticulum (ER) and the Golgi apparatus. The network of ER membranetubules aligns with microtubules and extends almost to the edge of the cell, whereas the Golgi apparatus is located near the centrosome. When cells are treated with a drug that depolymerizes microtubules, such as colchicine or nocodazole, the ER collapses to the center of the cell, while the Golgi apparatus fragments and disperses throughout the cytoplasm(). In vitro, kinesins can tether ER-derived membranes to preformed microtubule tracks, and walk toward the microtubule plus ends, dragging the ER membranes out into tubular protrusions and forming a membranous web very much like the ER in cells. Likewise, the outward movement of ER tubules toward the cell periphery is associated with microtubule growth in living cells. Conversely, dyneins are required for positioning the Golgi apparatus near the cell center, moving Golgi vesicles along microtubule tracks toward minus ends at the centrosome.
Effect of depolymerizing microtubules on the Golgi apparatus. (A) In this endothelial cell, the microtubules are labeled in red, and the Golgi apparatus is labeled in green (using an antibody against a Golgi protein). As long as the system of microtubules (more…)
The different tails and their associated light chains on specific motor proteins allow the motors to attach to their appropriate organelle cargo. For example, there is evidence for membrane-associated motor receptors, sorted to specific membrane-enclosed compartments, that interact directly or indirectly with the tails of the appropriate kinesin family members. One of these receptors seems to be the amyloid precursor protein, APP, which binds directly to a light chainon the tail of kinesin-I and is proposed to be a transmembrane motor proteinreceptormolecule in nerve-cell axons. It is the abnormal processing of this protein that gives rise to Alzheimer’s disease, as discussed in Chapter 15.
For dynein, attachment to membranes is known to be mediated by a large macromolecular assembly. Cytoplasmic dynein is itself a huge proteincomplex, and it requires association with a second large protein complex called dynactinto translocate organelles effectively. The dynactin complex includes a short actinlike filament that is made of the actin-related protein Arp1 (distinct from Arp2 and Arp3, the components of the ARP complex involved in the nucleation of conventional actin filaments). Membranes of the Golgi apparatus are coated with the proteins ankyrin and spectrin, which have been proposed to associate with the Arp1 filament in the dynactin complex to form a planar cytoskeletal array reminiscent of the erythrocyte membranecytoskeleton (see Figure 10-31). The spectrin array probably gives structural stability to the Golgi membrane, and—via the Arp1 filament—it may mediate the regulatable attachment of dynein to the organelle ().
A model for the attachment of dynein to a membrane-enclosed organelle. Dynein requires the presence of a large number of accessory proteins to associate with membrane-enclosed organelles. Dynactin is a large complex (red)that includes components that (more…)
Motor proteins also have a significant role in organelle transport along actin filaments. The first myosin shown to mediate organelle motility was myosin V, a two-headed myosin with a large step size (see ). In mice, mutations in the myosin V gene result in a “dilute” phenotype, in which fur color looks faded. In mice (and humans),membrane-enclosed pigment granules, called melanosomes, are synthesized in cells called melanocytes beneath the skin surface. These melanosomes move out to the ends of dendritic processes in the melanocytes, from where they are delivered to the overlying keratinocytes that form the skin and fur. Myosin V is associated with the surface of melanosomes, and it is able to mediate their actin-based movement in a test tube (). In dilute mutant mice, the melanosomes are not delivered to the keratinocytes efficiently, and pigmentation is defective. Other myosins, including myosin I, are associated with endosomes and a variety of other organelles.
Myosin V on melanosomes. (A) Phase-contrast image of a portion of a melanocyte isolated from a mouse. The black spots are melanosomes, which are membrane-enclosed organelles filled with the skin pigment melanin. (B) The same cell labeled with a fluorescent (more…)
Motor Protein Function Can Be Regulated
The cell can regulate the activity of motor proteins, allowing it to change either the positioning of its membrane-enclosed organelles or its whole-cell movements. One of the most dramatic examples is provided by fish melanocytes. These giant cells, which are responsible for rapid changes in skin coloration in several species of fish, contain large pigment granules that can alter their location in response to neuronal or hormonal stimulation (). These pigment granules aggregate or disperse by moving along an extensive network of microtubules. The minus ends of these microtubules are nucleated by the centrosome and are located in the center of the cell, while the plus ends are distributed around the cell periphery. The tracking of individual pigment granules reveals that the inward movement is rapid and smooth, while the outward movement is jerky, with frequent backward steps (). Both dynein and kinesinare associated with the pigment granules. The jerky outward movements apparently result from a tug-of-war between the two motor proteins, with the stronger kinesin winning out overall. When the kinesin light chains become phosphorylated after a hormonal stimulation that signals skin color change, kinesin is inactivated, leaving dynein free to drag the pigment granules rapidly toward the cell center, changing the fish’s color. In a similar way, the movement of other membrane organelles coated with particular motor proteins is controlled by a complex balance of competing signals that regulate both motor protein attachment and activity.
Regulated melanosome movements in fish pigment cells. These giant cells, which are responsible for changes in skin coloration in several species of fish, contain large pigment granules, or melanosomes (brown). The melanosomes can change their location (more…)
Bidirectional movement of a melanosome on a microtubule. An isolated melanosome (yellow) moves along a microtubule on a glass slide, from the plus end toward the minus end. Halfway through the video sequence, it abruptly switches direction and moves from (more…)
Myosin activity can also be regulated by phosphorylation. In nonmuscle cells, myosin II can be phosphorylated on a variety of sites on both heavy and light chains, affecting both motor activity and thick filament assembly. The myosin II can exist in two different conformational states in such cells, an extended state that is capable of forming bipolar filaments, and a bent state in which the tail domain apparently interacts with the motor head. Phosphorylation of the regulatory light chain by the calcium-dependent myosin light-chain kinase (MLCK) causes the myosin II to preferentially assume the extended state, which promotes its assembly into a bipolar filament and leads to cell contraction (). Myosin light-chain phosphorylation is an indirect target of activated Rho, the small GTPasediscussed previously whose activation causes a reorganization of the actincytoskeleton into contractile stress fibers. The MLCK is also activated during mitosis, causing myosin II to assemble into the contractile ring that is responsible for dividing the mitotic cell into two. Regulation of other members of the myosin superfamily is not as well understood, but control of these myosins is also likely to involve site-specific phosphorylations.
Light-chain phosphorylation and the regulation of the assembly of myosin II into thick filaments. (A) The controlled phosphorylation by the enzyme myosin light-chain kinase (MLCK) of one of the two light chains (the so-called regulatory light chain, shown (more…)
Summary
Motor proteins use the energy of ATP hydrolysis to move along microtubules or actin filaments. They mediate the sliding of filaments relative to one another and the transport of membrane-enclosed organelles along filament tracks. All known motor proteins that move on actin filaments are members of the myosin superfamily. The motor proteins that move on microtubules are members of either the kinesin superfamily or the dynein family. The myosin and kinesin superfamilies are diverse, with about 40 genes encoding each type of protein in humans. The only structural element shared among all members of each superfamily is the motor “head” domain. These heads can be attached to a wide variety of “tails,” which attach to different types of cargo and enable the various family members to perform different functions in the cell. Although myosin and kinesin walk along different tracks and use different mechanisms to produce force and movement by ATP hydrolysis, they share a common structural core, suggesting that they are derived from a common ancestor.
Two types of specialized motility structures in eucaryotic cells consist of highly ordered arrays of motor proteins that move on stabilized filament tracks. The myosin-actin system of the sarcomere powers the contraction of various types of muscle, including skeletal, smooth, and cardiac muscle. The dynein–microtubule system of the axoneme powers the beating of cilia and the undulations of flagella.
By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Copyright © 2002, Bruce Alberts, Alexander Johnson, Julian Lewis, Martin Raff, Keith Roberts, and Peter Walter; Copyright © 1983, 1989, 1994, Bruce Alberts, Dennis Bray, Julian Lewis, Martin Raff, Keith Roberts, and James D. Watson .
Bookshelf ID: NBK26888
Read Full Post »
2012pharmaceutical
sjwilliamspa
