Search

As more genome sequences are elucidated, there is an increasing need for information about the functions of the millions of proteins they encode. The function of a newly sequenced protein is often estimated by sequence alignment with the sequences of proteins with known functions. However, protein superfamilies can contain members that share significant amino acid sequence and structural homology yet catalyze different reactions or act on different substrates. Some homologous proteins differ by having a second or even third function, called moonlighting proteins. More recently, it was found that most protein superfamilies also include pseudoenzymes, a protein, or a domain within a protein, that has a three-dimensional fold that resembles a conventional catalytically active enzyme, but has no catalytic activity. In this review, we discuss several examples of protein families that contain enzymes, pseudoenzymes, and moonlighting proteins. It is becoming clear that pseudoenzymes and moonlighting proteins are widespread in the evolutionary tree, and in many protein families, and they are often very similar in sequence and structure to their monofunctional and catalytically active counterparts. A greater understanding is needed to clarify when similarities and differences in amino acid sequences and structures correspond to similarities and differences in biochemical functions and cellular roles. This information can help improve programs that identify protein functions from sequence or structure and assist in more accurate annotation of sequence and structural databases, as well as in our understanding of the broad diversity of protein functions.

In the cell, expression levels, allosteric modulators, post-translational modifications, sequestration, and other factors can affect the level of protein function. For moonlighting proteins, cellular factors like these can also affect the kind of protein function. This minireview discusses examples of moonlighting proteins that illustrate how a single protein can have different functions in different cell types, in different intracellular locations, or under varying cellular conditions. This variability in the kind of protein activity, added to the variability in the amount of protein activity, contributes to the difficulty in predicting the behavior of proteins in the cell.

Type I phosphomannose isomerases (PMIs) are zinc-dependent metalloenzymes involved in the reversible isomerization of D-mannose 6-phosphate (M6P) and D-fructose 6-phosphate (F6P). 5-Phospho-D-arabinonohydroxamic acid (5PAH), an inhibitor endowed with nanomolar affinity for yeast (Type I) and Pseudomonas aeruginosa (Type II) PMIs (Roux et al., Biochemistry 2004; 43:2926-2934), strongly inhibits human (Type I) PMI (for which we report an improved expression and purification procedure), as well as Escherichia coli (Type I) PMI. Its K(i) value of 41 nM for human PMI is the lowest value ever reported for an inhibitor of PMI. 5-Phospho-D-arabinonhydrazide, a neutral analogue of the reaction intermediate 1,2-cis-enediol, is about 15 times less efficient at inhibiting both enzymes, in accord with the anionic nature of the postulated high-energy reaction intermediate. Using the polarizable molecular mechanics, sum of interactions between fragments ab initio computed (SIBFA) procedure, computed structures of the complexes between Candida albicans (Type I) PMI and the cyclic substrate β-D-mannopyranose 6-phosphate (β-M6P) and between the enzyme and the high-energy intermediate analogue inhibitor 5PAH are reported. Their analysis allows us to identify clearly the nature of each individual active site amino acid and to formulate a hypothesis for the overall mechanism of the reaction catalyzed by Type I PMIs, that is, the ring-opening and isomerization steps, respectively. Following enzyme-catalyzed ring-opening of β-M6P by zinc-coordinated water and Gln111 ligands, Lys136 is identified as the probable catalytic base involved in proton transfer between the two carbon atoms C1 and C2 of the substrate D-mannose 6-phosphate.

Moonlighting proteins comprise a class of multifunctional proteins in which a single polypeptide chain performs multiple biochemical functions that are not due to gene fusions, multiple RNA splice variants or pleiotropic effects. The known moonlighting proteins perform a variety of diverse functions in many different cell types and species, and information about their structures and functions is scattered in many publications. We have constructed the manually curated, searchable, internet-based MoonProt Database (http://www.moonlightingproteins.org) with information about the over 200 proteins that have been experimentally verified to be moonlighting proteins. The availability of this organized information provides a more complete picture of what is currently known about moonlighting proteins. The database will also aid researchers in other fields, including determining the functions of genes identified in genome sequencing projects, interpreting data from proteomics projects and annotating protein sequence and structural databases. In addition, information about the structures and functions of moonlighting proteins can be helpful in understanding how novel protein functional sites evolved on an ancient protein scaffold, which can also help in the design of proteins with novel functions.

Phosphoglucose isomerase (PGI; E.C. 5.3.1.9) is a cytosolic enzyme that catalyzes the reversible isomerization of G6P and F6P. Its activity is important for glycolysis; gluconeogenesis; the pentose phosphate pathway; and the glycosylation of proteins, lipids, and other molecules. Recently, attention to PGI has grown because PGI is a moonlighting protein with a second function outside of the cell. PGI is the same protein as the extracellular growth factors neuroleukin (Chaput et al. 1988; Faik et al. 1988), autocrine motility factor (Watanabe et al. 1996), and differentiation and maturation mediator (Xu et al. 1996). It stimulates antibody secretion, promotes the survival of embryonic spinal neurons in vitro (Gurney et al. 1986a,b), affects an increase in tumor cell motility, and causes the differentiation of human myelogenous leukemia cells (Chiao et al. 1999) and a human myelocytic cell line (HL-60 cells) to terminal monocytic cells (Leung and Chiao 1985; Abolhassani et al. 1990). Recently, PGI also has been found to be the self-antigen in a mouse model of rheumatoid arthritis (Matsumoto et al. 1999; Schaller et al. 2001). One lab has identified antibodies to PGI in patients suffering from rheumatoid arthritis (Schaller et al. 2001). The enzymatic activity of PGI was first described over 70 years ago (Lohmann 1933). Biochemical characterization of PGI enzyme activity by several labs led to a proposed multistep mechanism involving general acid–base catalysis (Rose and O’Connell 1961; Rose 1975, 1981; Willem et al. 1990). Both G6P and F6P are substrates for PGI activity. Because G6P and F6P exist predominantly in their cyclic forms in solution (Swenson and Barker 1971), it is believed that first the cyclic form of G6P or F6P binds to the active site. PGI catalyzes ring opening of the substrate to yield the open chain form. The reaction then proceeds through the isomerization step using acid–base catalysis with proton transfer. PGI is a dimer with 557 amino acids in each subunit and has two equivalent active sites. X-ray crystal structures have been reported of PGI without bound ligand or with a substrate (F6P) or competitive inhibitors of catalytic activity bound in the active site pocket. X-ray crystal structures of mammalian (rabbit, pig, and human) PGI have been solved with 6PGA (Jeffery et al. 2000), 5PAA (Jeffery et al. 2001; Davies and Muirhead 2002; Davies et al. 2003), 5PAH (Arsenieva et al. 2002), or F6P (Lee et al. 2001), or with no bound ligand (Arsenieva and Jeffery 2002; Davies and Muirhead 2002, 2003). A crystal structure of human PGI with a sulfate ion was also reported (Read et al. 2001). Other crystal structures of PGI included two structures from a bacterium, Bacillus stearothermophilus, with no bound ligand (Sun et al. 1999) or complexed with 5PAA (Chou et al. 2000). Three crystal structures of PGI enzyme from Pyrococcus furiosus were also solved with no bound ligand (Berrisford et al. 2003), with 5PAA, and with 6PGA (Swan et al. 2003). Interestingly, the PGI enzyme from P. furiosus does not share sequence or structural homology with either mammalian or B. stearothermophilus PGI, and its activity is probably the result of convergent evolution. The X-ray crystal structures of PGI have helped to identify active site amino acids, ordered water molecules, and conformational changes involved in the ligand binding, ring opening, and isomerization steps of the multistep catalytic mechanism. Comparison of a crystal structure of rabbit PGI without a bound ligand to structures with ligands provided information about induced fit of the enzyme upon substrate binding. A crystal structure of rabbit PGI complexed with F6P provided information about the roles of Glu216, His388, Lys518, and an ordered water molecule in the ring opening step (Lee et al. 2001). Crystal structures of rabbit PGI complexed with 5PAA and 5PAH were used to identify the role of Glu357 as the active site amino acid residue that transfers a proton between C1 and C2 during the isomerization step (Jeffery et al. 2001) and the role of a water molecule in transferring a proton between the hydroxyl groups on C1 and C2 (Arsenieva et al. 2002). Arg272 was proposed to have a role in helping stabilize the cis-enedio-l(ate) intermediate. Comparison of the PGI/5PAA, PGI/5PAH, and PGI/F6P complexes also indicated that the substrate undergoes a rotation about the C3–C4 bond between the ring opening and isomerization steps to position C1 and C2 near to Glu357. Also, a helix containing amino acid residues 512–520 moves in toward the substrate to help hold it in place for the isomerization step. However, less clear is the state of the active site and substrate between the ring opening and isomerization steps. To complete the details of the catalytic mechanism of PGI, a structure of PGI complexed with an open chain substrate, or an analog, is required. Here, we report an X-ray crystal structure of rabbit PGI complexed with an analog of the open chain form of G6P (S6P), bound in the active site. This structure was solved at 2.0 A resolution.

Computer-based homology modeling techniques were used to construct a three-dimensional model of the Escherichia coli serine receptor ligand-binding domain based on the crystal structure of the Salmonella typhimurium aspartate receptor and the sequence homology between the two receptors. Residues that have been found in mutagenesis studies to be necessary for serine binding are located in a proposed serine-binding site. Several other mutations that affect swimming behavior require relatively small shifts in alpha-carbon positions in the model to give a minimized structure, suggesting that small changes in receptor conformation can affect the signaling state of the receptor.