Why is Moonlighting Important?
Many More Proteins Might Moonlight
The diverse examples of moonlighting proteins already identified suggest that many more moonlighting proteins are likely to found. The ability of one protein to perform multiple functions greatly expands the possible number of functions that can be performed by the proteome. In addition, the study of the molecular mechanisms and regulation of moonlighting functions helps broaden our understanding of protein biochemistry and suggests additional activities that might be encoded by genomes.
X-ray crystal structures and other biochemical and biophysical studies of some of the moonlighting proteins have added to our understanding of how one protein can perform two different functions and, in some cases, provided information about the triggers and molecular mechanisms involved in switching between two activities (several examples reviewed in Jeffery 2004, 2009).
In cases where the function of the protein changes in response to changes in the environment, moonlighting proteins provide examples of how a protein can sense and respond to these changes, and thereby provide interesting examples of regulation of protein function. The hemagglutinin-neuraminidase of paramyxovirus, which causes mumps, has different conformations at high- and low-pH conditions. The protein first enables binding of the virus to the surface of host cells. A change in pH promotes the movement of several amino acid side-chains and a loop in the active site to switch between the protein's sialic acid binding and hydrolysis functions so that it can cleave the glycosidic linkages of neuraminic acids (Crennell et al. 2000). The E. coli periplasmic serine endoprotease/heat- shock protein DegP (Protease Do) switches from a peptidase at high temperatures to a protein-folding chaperone at lower temperatures (Krojer et al. 2002).
In some moonlighting proteins the two functional sites are located distant from each other on the protein surface and the protein can perform both functions simultaneously, but in other proteins the functional sites are close to each other or even overlapping. Streptomyces coelicolor albaflavenone monooxygenase/syn- thase has a heme-dependent monooxygenase activity to catalyze the reaction (+)-epi-isozizaene + 2 NADPH + 2 O(2) < = > albaflavenone + 2 NADP(+) + 3 H(2) O and has a typical cytochrome P450 fold. However, the protein was also found to exhibit terpene synthase activity. After solution of its X-ray crystal structure a second active site pocket, for terpene synthase activity, was identified in an alpha- helical barrel near the monooxygenase active site (Zhao et al. 2008, 2009). It was recently found that the fructose-1,6-bisphosphate aldolase/phosphatase enzymes from the hyperthermophiles Thermoproteus neutrophilususes and Sulfolobus tokodaii utilize a single active site pocket to catalyze two reactions in the same biochemical pathway (Du et al. 2011; Fushinobu et al. 2011). After completion of the first catalytic function, several loops undergo conformational changes in order to bind the second substrate and perform the second catalysis.
A much larger conformational change occurs in cytosolic aconitase that renders the protein unable to perform one function but able to perform a second function. Aconitase is an enzyme in the citric acid cycle that uses an active site- bound 4Fe-4S cluster to catalyze the interconversion of citrate to isocitrate when cellular iron levels are high. When cellular iron concentrations decrease, the enzyme loses its 4Fe-4S cluster and becomes the iron-responsive element-binding protein (IRE-BP), which binds to iron-responsive elements (IREs) in 5'- or 3'- untranslated regions of mRNAs that encode proteins that are involved in iron uptake and use (Kennedy et al. 1992). This change in function involves a huge change in protein conformation. Domain 4 rotates 32° and translates 14 A relative to the rest of the protein, and domain 3 rotates 52° and translates 13 A relative to the rest of the protein. All four protein domains then interact with the IREs. In fact, the RNA binding and active sites overlap extensively, and several conserved active site amino acids are also important in mRNA binding (Philpott et al. 1994; Walden et al. 2006).
Other moonlighting proteins perform one function as a monomer or homo- multimer but are incorporated into a larger multiprotein complex (i.e., ribosome, proteasome) to perform a second, sometimes structural role, and in those cases the moonlighting polypeptide might or might not undergo a large conformational change. The ferredoxin-dependent glutamate synthase in spinach chloro- plasts (FdGOGAT) catalyzes the reaction 2 L-glutamate + 2 oxidized ferredoxin = > L-glutamine + 2-oxoglutarate + 2 reduced ferredoxin in L-glutamate biosynthesis and is also a subunit of the UDP-sulfoquinovose synthase (SQD1) (Shimojima et al. 2005), a multiprotein complex that catalyzes the transfer of sulfite to UDP-glucose in the synthesis of UDP-sulfoquinovose, which is the head group donor in the biosynthesis of sulfoquinovosyldiacylglyerol, a plant sulfolipid. Human delta-aminolevulinic acid dehydratase (porphobilinogen synthase, ALADH) converts 5-aminolevulinate to porphobilinogen in the biosynthesis of protoporphyrin-IX (Gibbs and Jordan 1986; Wetmur et al. 1986). It also interacts with the proteasome as a proteasome inhibitory subunit that blocks proteolysis of specific protein substrates (Li et al. 1991; Guo et al. 1994). Tetrahymena thermophile citrate synthase, a soluble enzyme from the citric acid cycle, is a protein in the 14 nm cytoskeletal filament (Numata 1996).
Questions remain as to how these proteins obtained a second function. Many of the known moonlighting proteins are ubiquitous enzymes in central metabolism or ubiquitous chaperone proteins (Jeffery 1999, 2009). These proteins first arose billions of years ago and are expressed in many species and cell types. They are likely to have been adopted for a second function because organisms evolve by utilizing and building upon components they already possess, and these proteins are available in many organisms.
Binding to another protein is the key characteristic of the second, or more recently acquired, function of many of the known moonlighting proteins, and a new binding function can result if a protein's structure is modified to create a new binding site on the protein surface. Modification of a short amino acid sequence on a surface exposed loop could be all that is needed for the formation of a new protein-protein interaction site. Enolase is an ubiquitous cytoplasmic enzyme in glycolysis that moonlights in many species. In several bacterial species it is found on the cell surface where it can bind to plasminogen. The plasminogen-binding site of Streptococcus pneumoniae enolase has been identified as a nonapeptide (248-FYDKERKVY-256). In X-ray crystal structures of S. pneumoniae enolase (PDB-ID = 1W6T), this sequence motif is found to be on the solvent exposed surface of the octamer, in surface loop near the active site pocket. Three loops, L1 (residues 38-45), L2 (152-159), and L3 (244-265), fold over the active site pocket in the substrate-bound state. The plasminogen binding site is located in L3 (Bergmann et al. 2003; Ehinger et al. 2004).
In general, enzymes appear to contain many more amino acids than are required to form an active site pocket, leaving a lot of surface amino acids that are not involved in the original function and are therefore not under as much selective pressure. This has been illustrated by considering the X-ray crystal structure of phosphoglucose isomerase, an enzyme that is nearly ubiquitous in evolution (Jeffery et al. 2000). In an alignment of 136 PGI sequences, 47 residues were found to be conserved. The conserved amino acids include those that interact with substrate as well as others that help form the shape of the active site pocket and position the catalytic amino acids, and these are the amino acids that have been conserved for three billion years of evolution to maintain the isomer- ase activity. At the same time, during three billion years of evolution most of the other residues changed. As is the case with most proteins, many of the amino acids that have undergone changes are located on the protein's surface. In this large dimeric protein of more than 1000 amino acids, numerous entire helices and other surface features are made up of nonconserved residues. It is quite possible that one of these surface features could have gained an additional binding function during evolution. As long as that new function did not interfere with the isomerase activity of the protein, it had the potential to benefit the organism and its offspring and was perhaps kept during evolution. This is one possible way to explain how a protein can evolve a moonlighting function.
Additional insight into how a protein can gain a novel function is provided by the identification of several “neomorphic moonlighting proteins,” proteins that gain a second function through a single amino acid substitution (reviewed in Jeffery 2011). These proteins are not true moonlighting proteins because a second function is performed only by a mutant form of the protein and is not a normal physiological function of the protein. Several of these gain-of-function mutations have been identified because they result in disease.
Dihydrolipoamide dehydrogenase (DLD) is a flavin-dependent oxidoreductase that is found in several multienzyme complexes, including the pyruvate, alpha- ketoglutarate, and branched chain amino acid dehydrogenase complexes. Wild- type DLD is a dimer that catalyzes the conversion of dihydrolipoic acid to lipoic acid along with the reduction of NAD+ to NADH. Because of its critical role in energy and redox balance in the cell, genetic mutations that cause a deficiency of enzyme activity result in severe disorders in infancy; the symptoms are however variable and due to the specific mutation found in each case. Some single amino acid substitutions in the homodimer interface result in hypertrophic cardiomyopathy, which is not observed in patients with other mutations in the protein. Surprisingly, these mutations cause a decrease in dimer formation and reveal a protease active site that enables the enzyme to cleave protein substrates, which might contribute to the observed symptoms (Babady et al. 2007; Brautigamet et al. 2005). The neomorphic moonlighting proteolytic activity was shown to be independent of the DLD activity because a S456A amino acid substitution, which is in the catalytic dyad of the protease active site, abolished protease activity but did not affect DLD activity.
Mutations in isocitrate dehydrogenases result in a novel product of catalysis that promotes cancer (Yan et al. 2009; Figueroa et al. 2010). The NADP + - dependent isocitrate dehydrogenases (IDH1 and IDH2) catalyze the oxidative decarboxylation of isocitrate to alpha-ketoglutarate (alpha-KG) in the Krebs cycle. The cause of some gliomas and some cases of acute myeloid leukemia (AML) was found to be an amino acid substitution at R132 in the catalytic pocket of IDH1. In other gliomas, substitutions were found at the equivalent R172 position in IDH2. Instead of knocking out enzyme activity, the single amino acid substitution mutation causes a neomorphic moonlighting enzymatic activity. In place of producing the usual alpha-ketoglutarate product, the mutant enzyme reduces alpha-KG to the R-enantiomer of 2-hydroxyglutartate ((R)-2HG R(2)-2-hydroxyglutarate; 2HG; Dang et al. 2010; Xu et al. 2011; Lu et al. 2012). The 2HG product is an oncometabolite that works by inhibiting alpha-KG-dependent dioxygenases, including proteins involved in histone and DNA demethylation, thereby affecting the epigenetic state of the cells and blocking cellular differentiation.
The fact that single amino acid substitutions in DLD and IDH can cause a gain of function (although a pathological “neomorphic” moonlighting function and not a “true” moonlighting function) that results in changes in the cell suggests that, at least in some cases, very small changes in a protein sequence and structure might be all that is needed for a protein to gain a true moonlighting function.
It is also interesting to note that an ancestral protein can gain different moonlighting proteins in different evolutionary lineages. Enolase and GAPDH are enzymes and plasminogen or extracellular matrix-binding proteins in several species, as described above. In the sea lamprey (Petromyzon marinus), enolase is an enzyme and has been adopted to be tau-crystallin in the lens of the eye (Stapel and de Jong 1983; Williams et al. 1985; Jaffe and Horwitz 1992). In the diurnal gecko (Phelsuma), GAPDH is an enzyme and also pi-crystallin (Jimenez-Asensio et al. 1995). The protein-folding chaperonin 60 has also been adopted for many different moonlighting functions in different species (Henderson et al. 2013).
Further discussion of the current knowledge of evolution of protein moonlighting and its structural biological underpinnings are the topics of Chapters 2 and 3.