The original concerted and sequential models of allostery, promoting the view of stable oligomers, remain the most often considered. Since 1970, the relationship between dissociation/association and allostery has been rarely invoked; there have been only five reviews [75
], the most recent of which was more than fifteen years ago, and two books [79
], which consider this relationship. However, throughout the past decade we have characterized a protein for which allostery involves oligomer dissociation, conformational change in the dissociated state, and reassembly to a structurally and functionally distinct oligomer [7
]. Our initial unwillingness to embrace the possibility for this particular type of quaternary structure dynamic was seated in our general appreciation for the one sequence → one structure paradigm, which we, like most, considered to be applicable on the level of quaternary structure. This enlightening system, which is described in detail in an accompanying article [33
], is introduced below along with consideration of whether other proteins may behave in a similar fashion. The homooligomeric protein is porphobilinogen synthase (PBGS), which catalyzes the first common step in the synthesis of the tetrapyrrole pigments (heme, chlorophyll, vitamin B12
). PBGS can exist as a high activity octamer that can dissociate to dimers; the dimers can populate alternate conformations; and one alternate dimer conformation can assemble into a low activity hexamer () [34
]. In some species there is an allosteric activator that stabilizes a subunit interface present in the octamer and absent in the hexamer [34
]. These findings led to a new dissociating model for allosteric regulation termed the morpheein model [7
]; this is described in more detail below.
Morpheein model of allosteric regulation
The morpheein model of allosteric regulation [7
] marries conformational diversity with quaternary structure dynamics and establishes a physiological relevance to the pressure induced quaternary structure rearrangements described as conformational drift. In the context of the morpheein model, homo-oligomer assemblies readily associate and dissociate under physiologic conditions and there is sufficient conformational flexibility in the dissociated state to dictate formation of more than one kind of oligomeric assembly. A morpheein is defined to be a protein that exists as an equilibrium of quaternary structure assemblies such as is illustrated in by both a three-dimensional dice analogy () and a two-dimensional geometric () analogy. The assembled higher order oligomers are of finite multiplicity and different functionality. The individual states (morpheein forms) are linked by dynamic equilibria such that interchange between higher order oligomers must involve dissociation. The dissociated form has conformational flexibility and the specific conformation of this form determines which higher order oligomer is assembled. The change in conformation involves a small number of peptide backbone angles, such as hinge motion between domains; it does not involve protein unfolding/refolding. Assembly of oligomers is such that symmetrical contacts are maintained between the subunits in each oligomer (one dot to four dots for the dice; dashed line to thick line for the geometric shapes). The morpheein forms are all relatively close in energy, e.g. the low population states are present at some, albeit small, mole fraction. In this way the interconversion of morpheein forms can be accomplished under normal physiologic (native) conditions. The equilibrium can be drawn in one direction or the other through non-covalent ligand binding (); such allosteric ligands exert their effect on protein function by stabilizing one or another of the functionally distinct morpheein forms.
In the morpheein allosteric model, all forms are accessible in the presence and absence of allosteric effector molecules; the dynamic equilibria are merely shifted through binding of effectors. This is emphasized in by the position of the equilibria in the presence or absence of appropriately shaped allosteric effector molecules. While stable alternate conformations (or quaternary assemblies) are often associated with post-translational modifications, interchange between morpheein forms occurs without such modification. Without chemical modification, the ready interchange of morpheein forms demands that they be of similar energy, as has been demonstrated for the morpheein forms of PBGS [34
]. There may however be a substantial kinetic barrier to the interconversion process.
What rules might dictate how identical subunits can rearrange?
In general, protein subunits assemble with themselves (homomeric assembly) such that they retain an element of symmetry wherein equivalent surfaces of equivalent monomers are in equivalent environments (see ). If one monomer has, for instance, residue X in a hydrogen bond with residue Y of another subunit, then it is preferred that all residues X of each subunit are in a hydrogen bond with all residues Y of a neighboring subunit. Complete dissociation of an oligomer () releases all these interactions simultaneously, retaining the equivalence of the protein surfaces, one subunit to the next. But removal of one subunit from the trimer or tetramer would result in a loss of this equivalence and establish an asymmetric structure. In a given environment (pH, temperature, ionic strength), the physical chemical forces governing association of the subunits are equivalent for all subunits and the system will strive to retain symmetry. Thus, there are many oligomeric assemblies larger than a dimer where partial dissociation creates asymmetry. For the PBGS octameric assembly, illustrated in , loss of one subunit or one dimer would result in such asymmetry; splitting the octamer in half by removing two pro-octamer dimers also results in asymmetry. However, separating the octamer into two halves in a process that cleaves each pro-octamer dimer (along the barrel-to-barrel interface) would retain symmetry. The resulting tetramers would each contain four dangling N-terminal arms. Although there are valid arguments against this latter example, such as the significance of the pulled apart arms to the structural integrity barrel-barrel contacts, symmetry is retained. However, one way that PBGS could dissociate and still retain both symmetry and two out of the three surface-to-surface contacts would be to simultaneously dissociate to four dimmers (). As described above, it is the need to maintain symmetry and the orientation of the Nterminal arm that drives formation of the hexamer and/or octamer of PBGS. For an in depth discussion of oligomeric assembly, the reader is referred to a review article by Goodsell and Olson [90
Generation of protein asymmetry during multimer disassembly
Morpheeins expand our understanding of disease and drug action
Inborn errors of metabolism often result from mutations that alter the function of essential proteins. If the natural function of a particular protein relies on its ability to equilibrate between morpheein forms, then single amino acid variations that alter the equilibrium of such forms can result in a disregulation of said protein function and cause or contribute to a disease state. This is indeed the case for human PBGS in relation to the disease of ALAD porphyria; all of the naturally occurring mutations of human PBGS that are associated with the disease ALAD porphyria have been shown to shift the quaternary structural equlibria toward the low activity hexameric assembly [83
]. Such diseases can be considered among the conformational diseases. Hence, identification of proteins that function as morpheeins can improve our understanding of the structural basis of human diseases associated with the function of these protein and/or variations in disease susceptibility. Several of the putative morpheeins in are associated with inborn errors of metabolism. These are 1) alpha-galactosidaseA, which is associated with Fabry disease, 2) the ATPase of the ABCA1 transporter, which is mutated in Tangier’s disease and familial high-density lipoprotein deficiency, 3) cystathionine-β-synthase, for which some mutations cause a form of homocystinuria, 4) dihydrolipoamide dehydrogenase, associated with E3-deficient maple syrup urine disease and lipoamide dehydrogenase deficiency, 5) NAD(+)-dependent mitochondrial malic enzyme, where mutations are related to certain hereditary forms of epilepsy, 6) phenylalanine hydroxylase deficiency, which causes phenylketonuria, and 7) pyruvate kinase, whose deficiency is the most common cause of hemolytic anemia. For each of these proteins, there are one or more publications that justify a model of functional control via equilibrium of alternate protein assemblies (see ).
Just as understanding disease mechanism may be improved by considering that certain proteins function as morpheeins, so too can our understanding of the mechanism of action of effective therapeutics whose mechanism is not otherwise explained. In proof of this concept small molecule allosteric inhibitors of PBGS, which function by stabilizing the inactive hexameric assembly, have been identified [84
]. Following the identification of the morpheein character of medically relevant proteins, comes the possibility of designing or discovering allosteric inhibitors or activators of these proteins. The literature suggests that small molecule allosteric stabilization of alternative quaternary structures may be the an effective approach to finding inhibitors directed against prominent drug targets such as tumor necrosis factor α [91
] or HIV integrase [92
]. To this end the characteristics of a morpheein as these may aid in the identification of such proteins are discussed.
Characteristics that suggest a protein functions as a morpheein
There are myriad characteristics that suggest, but do not prove, that a particular protein may function as a morpheein. provides a list of proteins that exhibit one or more of these characteristics and are thus suspect of being able to exist as an equilibrium of functionally distinct, non-additive quaternary structure assemblies or morpheein forms. Each entry in includes at least one reference which the reader can use to begin exploring the putative morpheein nature of this protein. Some suggestive characteristics are illustrated in .
Some characteristics of a morpheein ensemble
Protein concentration dependent specific activity ()
Specific activity (units of enzyme activity per protein mass) is a value often used during protein purification to assess protein purity. In the case of monomeric enzymes, or stable enzyme oligomers, specific activity is expected to be independent of the concentration of the enzyme used in the assay. However, for an equilibrium of morpheeins forms, if rapid interchange of oligomeric assemblies is occurring during the assay, the observed specific activity may be dependent upon the concentration of the enzyme in the assay mixture. This relationship derives from the law of mass action, which favors the largest oligomer as the protein concentration is increased in the range of its inherent dissociation constant. The illustration in refers to a situation where the largest oligomer is the most active. In the case where a smaller oligomer is most active, there will be an inverse relationship between protein concentration and specific activity, as has been reported for purine nucleoside phosphorylase [93
]. However, it is important to note that one may only recognize such a protein concentration dependent specific activity if one is varying the protein concentration within the range (+/− one order of magnitude) of the apparent K0.5
. Outside that range the effect will be too small to discriminate from experimental error.
One is classically taught that initial rates of enzyme catalyzed reactions should display a linear increase in the concentration of product. However, in the case of an enzyme that functions as a morpheein, this may not be so. Upon dilution into assay buffer, a lag (shown in ) or a burst in activity may be apparent as the equilibrium of morpheein forms adjusts to assay conditions. In the example in the assay conditions promote a shift to a more active morpheein form and the rate of product production increases with time. A shift to a less active form would have the opposite with a decrease in the rate of product production occurring with time. The term hysteresis was coined by Frieden in 1970 to describe non-linear kinetics [94
]. Subsequently Kurganov presented mathematical models to explain various hysteretic behavior [79
]. While the early work on kinetic hysteresis did not have the luxury of the structural insight gained from a decade of work on the PBGS quaternary structure equilibrium, one explanation for hysteretic behavior is a switch in the distribution of morpheein forms. The work of Frieden and Kurganov also points to the fact that kinetic hysteresis is a possible characteristic of an enzyme that functions as a morpheein, but it is not a diagnostic tool by itself.
Non-traditional Michaelis Menten kinetics
The existence of multiple forms of an enzyme may be reflected by non-traditional Michaelis-Menten plots. Generally it is expected that a Michaelis-Menten plot will be hyperbolic unless there is some phenomenon such as product inhibition to perturb this trend. In the case of a morpheein it is possible that two or more catalytically active forms may be present. Provided that the affinities of the various forms for the substrate are sufficiently different, a nontraditional Michaelis-Menten plot will be produced that consists of more than one hyperbola, the number of which corresponds to the number of active forms present in the assay. For example, the predominant forms of human PBGS are the octamer and the hexamer. These two forms have vastly different Km
values and the result is a double hyperbolic relationship between substrate concentration and activity. However, to see this relationship, it is necessary that experimental substrate concentrations cover the range of the two Km
Multiple quaternary structures
Often multiple quaternary structural forms that are visualized by gel filtration chromatography or native-PAGE are viewed as representing an impure sample. This is especially true if one or more of the separated forms are found to be inactive. However, if a protein functions as a morpheein, the multiple quaternary structures may be functionally distinct forms of the same protein. Rechromatography under different conditions may produce a different distribution of quaternary structural forms and even more confounding is the fact that these forms may have vastly different functions (see moonlighting proteins below). In addition to separation by traditional sizing methods the surface charge differences between the different quaternary structural forms may allow separation by ion-exchange chromatography. For example, the octameric, hexameric, and dimeric forms of PBGS can be separated by anion exchange chromatography [34
pH profile effects
pH profiles also may depend on the conditions under which they are obtained. For example, the alkaline limb of the human PBGS pH-activity profile is due at least in part to the pH dependence of the quaternary structure equilibrium. High pH favors the less active hexamer compared to the more active octamer [85
Order of addition effects
Incubation conditions prior to addition of the reaction initiating compound can influence the distribution of morpheein forms of a protein. GDP-mannose dehydrogenase (GMD) provides a striking example of order of addition effects [95
]. The substrates for GMD are GDP-mannose and NAD+
, and both phosphate and GMP act as allosteric effectors [95
]. As shown in the four very different reaction rate curves are produced when the assays were carried out using four different orders of addition of reactants and/or allosteric effectors. Similar phenomena have also been referred to as enzyme memory [96
It has become common to find that one protein has multiple unrelated functions; these proteins are said to moonlight [97
]. One notable example is the glycolytic enzyme glyceraldehydes-3-phosphate dehydrogenase which has been reported to have many activities unrelated to glycolysis [98
–104]. It is possible that proteins that moonlight may not always do so as the same quaternary structure assembly. It is not clear if glyceraldehyde-3-phosphate dehydrogenase (GAPDH) performs all of its activities as the tetramer believed to be functional in glycolysis. It may be possible to modulate a preferred function of multimeric moonlighting proteins with small molecules that modulate an equilibrium of morpheein forms.
Morpheeins in relation to other oligomerizing proteins
Morpheeins represent a subset of oligomerizing proteins. The features that distinguish morpheeins from other oligomerizing proteins are that they are homo-oligomeric, of finite stoichiometry, must dissociate/associate to interchange between functionally distinct multimers, and assembly of oligomers is autonomous, i.e., that is oligomer assembly occurs without the need for covalent modification or the presence of a template/scaffold. We have chosen to limit the definition of morpheeins to homo-oligomeric proteins because the archetypical morpheein PBGS is such a protein. There is no direct evidence for or against hetero-oligomers existing in dynamic equilibria such that they dissociate, change shape and come back together differently to regulate function. However, on the simplest level hetero-oligomers can be viewed as one protein being regulated by interaction with another; a scenario that does not require morpheein-like behavior.
Among homo-oligomeric proteins, morpheeins represent a distinct category. Conformational flexibility is evident in many homo-oligomers that do not follow the morpheein model. For example, prions adopt at least two conformations; one that does not oligomerize and one that does [105
]. In terms of the morpheein model these could be viewed as conformations of a fundamental unit. However, oligomerization of prions differs markedly from that of morpheeins. The symmetry of oligomerization does not dictate a size limit for the oligomer such that large fibrils may be formed. Furthermore the thermodynamics of fibril formation is generally not amenable to a physiologically relevant and readily reversible mechanism for allostery.
Virus coat proteins also bear similarities to, but are not, morpheeins. These proteins must assemble to form a capsid around the virus genome to protect it from damage but upon invading a cell the coat must disassemble to allow infection by release of the genome. These properties imply that there are two conformations of the protein only one of which is capable of oligomerization. To form the capsid the proteins assemble with either a helical or an icosahedral symmetry in which the subunits may lie in quasi-equivalent environments [106
]. However, the requirement to form a stable barrier necessitates that the oligomer is not in dynamic equilibrium, but that the oligomer forming conformation is a metastable state that remains until some trigger removes a barrier to a conformational change that favors dissociation. Further distinctions are that, this conformational change must occur in the oligomer, not all capsids are homo-oligomeric, and the genome or some scaffold protein generally forms a template for the capsid.