Molybdenum is the only second-row transition metal required by most living organisms, and is nearly universally distributed in biology. Enzymes containing molybdenum in their active sites have long been recognized,1 and at present over 50 molybdenum-containing enzymes have been purified and biochemically characterized; a great many more gene products have been annotated as putative molybdenum-containing proteins on the basis of genomic and bioinformatic analysis.2 In certain cases, our understanding of the relationship between enzyme structure and function is such that we can speak with confidence as to the detailed nature of the reaction mechanism and, with the availability of high-resolution X-ray crystal structures, the specific means by which transition states are stabilized and reaction rate is accelerated within the friendly confines of the active site. At the same time, our understanding of the biosynthesis of the organic cofactor that accompanies molybdenum (variously called molybdopterin or pyranopterin), the manner in which molybdenum is incorporated into it, and then further modified as necessary prior to insertion into apoprotein has also (in at least some cases) become increasingly well understood.

It is now well-established that all molybdenum-containing enzymes other than nitrogenase (in which molybdenum is incorporated into a [MoFe7S9] cluster of the active site) fall into three large and mutually exclusive families, as exemplified by the enzymes xanthine oxidase, sulfite oxidase, and DMSO reductase; these enzymes represent the focus of the present account.3 The structures of the three canonical molybdenum centers in their oxidized Mo(VI) states are shown in Figure 1, along with that for the pyranopterin cofactor. The active sites of members of the xanthine oxidase family have an LMoVIOS-(OH) structure with a square-pyramidal coordination geometry. The apical ligand is a Mo=O ligand, and the equatorial plane has two sulfurs from the enedithiolate side chain of the pyranopterin cofactor, a catalytically labile Mo–OH group, and most frequently a Mo=S. Nonfunctional forms of these enzymes exist in which the equatorial Mo=S is replaced with a second Mo=O; in at least one member of the family the Mo=S is replaced by a Mo=Se, and in others it is replaced by a more complex –S–Cu–S–Cys to give a binuclear center. Members of the sulfite oxidase family have a related LMoVIO2(S–Cys) active site, again square-pyramidal with an apical Mo=O and a bidentate enedithiolate Ligand (L) in the equatorial plane but with a second equatorial Mo=O (rather than Mo–OH) and a cysteine ligand contributed by the protein (rather than a Mo=S) completing the molybdenum coordination sphere. The final family is the most diverse structurally, although all members possess two (rather than just one) equiv of the pyranopterin cofactor and have an L2MoVIY(X) trigonal prismatic coordination geometry. DMSO reductase itself has a catalytically labile Mo=O as Y and a serinate ligand as X completing the metal coordination sphere of oxidized enzyme. Other family members have cysteine (the bacterial Nap periplasmic nitrate reductases), selenocysteine (formate dehydrogenase H), –OH (arsenite oxidase), or aspartate (the NarGHI dissimilatory nitrate reductases) in place of the serine. Some enzymes have S or even Se in place of the Mo=O group. Members of the DMSO reductase family exhibit a general structural homology to members of the aldehyde:ferredoxin oxidoreductase family of tungsten-containing enzymes;4 indeed, the first pyranopterin-containing enzyme to be crystallographically characterized was the tungsten-containing aldehyde:ferredoxin oxidoreductase from Pyrococcus furiosus,5 a fact accounting for why many workers in the field prefer “pyranopterin” (or, perhaps waggishly, “tungstopterin”) to “molybdopterin”. The term pyranopterin will generally be used in the present account.




Open in a separate window


Figure 1


Active site structures for the three families of mononuclear molybdenum enzymes. The structures shown are, from left to right, for xanthine oxidase, sulfite oxidase, and DMSO reductase. The structure of the pyranopterin cofactor common to all of these enzymes (as well as the tungsten-containing enzymes) is given at the bottom.

/pdf/the-mononuclear-molybdenum-enzymes-huemv59go8.pdf

The Mononuclear Molybdenum Enzymes

Site-specific ribose methylation of preribosomal RNA: a novel function for small nucleolar RNAs.

In an attempt to create an animal model of tissue-specif ic mitochondrial disease, we generated ‘knockout’ mice deficient in the heart/muscle isoform of the adenine nucleotide translocator (Ant1). Histological and ultrastructural examination of skeletal muscle from Ant1 null mutants revealed ragged-red muscle fibers and a dramatic proliferation of mitochondria, while examination of the heart revealed cardiac hypertrophy with mitochondrial proliferation. Mitochondria isolated from mutant skeletal muscle exhibited a severe defect in coupled respiration. Ant1 mutant adults also had a resting serum lactate level fourfold higher than that of controls, indicative of metabolic acidosis. Significantly, mutant adults manifested severe exercise intolerance. Therefore, Ant1 mutant mice have the biochemical, histological, metabolic and physiological characteristics of mitochondrial myopathy and cardiomyopathy.

A mouse model for mitochondrial myopathy and cardiomyopathy resulting from a deficiency in the heart/muscle isoform of the adenine nucleotide translocator

In the nucleolus of eukaryotic cells, ribosomal DNA is transcribed by RNA polymerase I into long precursor (pre-rRNA) transcripts, which are modified by methylation and pseudo-uridylation, cleaved to yield 18S, 5.8S, and 28S rRNAs, and then assembled into the mature large and small ribosomal subunits prior to export to the cytoplasm (for reviews see references 22, 26, and 77). A large number of small nucleolar ribonucleoprotein (snoRNP) particles have emerged as key players in this biosynthetic process. Currently more than 70 snoRNA species have been identified (for reviews see references 51, 75, and 82). All snoRNAs, with the exception of MRP RNA, can be divided into two classes: those that possess boxes C (RUGAUGA) and D (CUGA), which are required for association with the abundant nucleolar autoantigen fibrillarin (see reference 51), and those that possess boxes H (ANANNA) and ACA, which mediate the binding of Gar1 protein (4, 8, 24, 37). Only a few snoRNAs have been found to be required for growth in yeast (U3 [6, 29], U14 [43], MRP [14, 73], snR10 [80, 81], and snR30 [5, 54]) or for specific pre-rRNA cleavage events in Xenopus oocytes (U3 [34, 72], U8 [63], and U22 [85]).

Recently, box C/D snoRNAs and box H/ACA snoRNAs were found to target specific sites in pre-rRNA for 2′-O-methylation and pseudouridylation, respectively (for reviews see references 46, 48, 62, 75, and 82). These modification reactions are mediated by extensive regions (10 to 21 nt) of complementarity between the so-called “antisense” snoRNAs and sequences flanking the rRNA sites to be modified. Specifically, U24, U20, and U25 were shown to direct site-specific ribose methylation of pre-rRNA in HeLa cells (39), yeast (13), and Xenopus oocytes (87), respectively, and snR8, snR3, snR33, and snR5 (among others) were demonstrated to target pre-rRNA for pseudouridylation in yeast (23, 57). The presence of ∼200 modified nucleotides in vertebrate rRNA (47) suggests that more than half of the antisense snoRNAs remain to be identified.

A unique feature of snoRNAs is that most are encoded within the introns of protein-coding genes (reviewed in reference 51). This economic use of introns is commonplace among intron-rich organisms, such as vertebrates, where antisense snoRNAs have been found to be exclusively intron encoded. By contrast, in Saccharomyces cerevisiae many snoRNAs are produced from independent transcription units. A vertebrate host gene intron encodes only a single snoRNA, whereas in yeast and plants, some snoRNA genes are located in polycistronic arrays without exons separating the snoRNA sequences (41). In all species investigated, the intron-encoded snoRNAs are transcribed from their host genes by RNA polymerase II as portions of the pre-mRNA. The functional snoRNAs are then produced by exonucleolytic trimming that follows either splicing (12, 38, 86) or endonucleolytic cleavage of intron sequences (9, 10).

The mode by which snoRNA sequences became inserted into the introns of their host genes is not known. Interestingly, the host gene for a particular snoRNA can differ even among closely related vertebrates, suggesting that intron-encoded snoRNAs may be highly mobile genetic elements (see reference 51). Likewise, the reason particular genes have been chosen as hosts for intron-encoded snoRNAs has been unclear. Initially, it appeared that all snoRNA host genes generate protein products that function in ribosome biogenesis or in translation; ribosomal proteins (rp) L1, L5, L7a, S8, nucleolin, and eIF4AI are a few examples (see references 66, 68, 67, 59, 58, and 24, respectively). Such genetic organization could provide coregulation of protein components of the translational machinery and snoRNAs, which contribute to rRNA maturation (76). However, the discovery of other snoRNA host genes lacking obvious ribosome-related functions (for example, ATP synthase β [39]) suggested that host genes may have been chosen merely to meet the need for transcription rates high enough to produce a sufficient level of snoRNAs (∼104 copies/cell) to base pair with the cell’s nascent pre-rRNA molecules.

UHG (U22 host gene) is an unusual snoRNA host gene because it does not appear to specify a protein product. It generates, in addition to U22, seven different box C/D antisense snoRNAs (U25 to U31) (84). Comparison of mouse and human UHG sequences revealed that its introns are more conserved than its exons, suggesting that the snoRNAs may be the only functional portions of the transcript. Both human and mouse UHG messages are riddled with stop codons in all three reading frames; the longest open reading frames (ORFs) would produce peptides of 51 and 40 amino acids for human and mouse UHG, respectively. Nonetheless, the UHG transcript resembles a typical rp mRNA in that it begins with a C residue, is spliced and polyadenylated, and is associated with ribosomes. However, unlike rp mRNAs, the spliced UHG RNA is almost undetectable in HeLa cells. Inhibition of translation in HeLa cells with the initiation inhibitor pactamycin or elongation inhibitors cycloheximide or puromycin results in a 15-fold increase in the level of spliced UHG transcript (84). This link between the levels of UHG RNA and active translation, in conjunction with its numerous stop codons, suggests that it may be a candidate for the nonsense-mediated decay pathway (49, 84).

Here we report the identification of a second member of the UHG class of snoRNA host genes. Growth arrest-specific transcript 5 (gas5) was initially discovered in a screen for potential tumor suppressor genes expressed at high levels during growth arrest (74). The murine gas5 gene produces a ubiquitous, polyadenylated, alternatively spliced message which is almost undetectable in actively growing cells yet is highly expressed in cells undergoing serum starvation or density arrest (15, 16). We demonstrate that gas5 is a multi-snoRNA host gene which encodes 9 (in mouse) or 10 (in human) antisense snoRNAs. By mapping the 5′ end of the gas5 transcript and comparing it with other known snoRNA host genes, we observe that all known snoRNA host genes exhibit characteristics which define the 5′TOP (terminal oligopyrimidine) class of genes. We provide evidence that membership in the 5′TOP family explains why the abundance of the gas5 spliced product is growth dependent. Furthermore, the discovery that all snoRNA host genes contain 5′TOP sequences may illuminate why certain genes have been selected to serve as snoRNA host genes.

Classification of gas5 as a multi-small-nucleolar-RNA (snoRNA) host gene and a member of the 5'-terminal oligopyrimidine gene family reveals common features of snoRNA host genes.

Facioscapulohumeral muscular dystrophy (FSHD) is a dominantly inherited disorder with an initially restricted pattern of weakness. Early involvement of the facial and scapular stabilizer muscles results in a distinctive clinical presentation. Progression is descending, with subsequent involvement of either the distal anterior leg or hip-girdle muscles. There is wide variability in age at onset, disease severity, and side-to-side symmetry, which is evident even within affected members of the same family. Although FSHD is considered a relatively benign dystrophy by some, as many as 20% of patients eventually become wheelchair-bound. Associated nonskeletal muscle manifestations include high-frequency hearing loss as well as retinal telangiectasias, both of which are rarely symptomatic. The causal genetic lesion in FSHD was described over a decade ago, raising hope that knowledge about its molecular and cellular pathophysiology was soon to follow. In the vast majority of cases, FSHD results from a heterozygous partial deletion of a critical number of repetitive elements (D4Z4) on chromosome 4q35; yet, to date, no causal gene has been identified. The accumulating evidence points to a complex, perhaps unique, molecular genetic mechanism. The absence of detectable expressed sequences from D4Z4, the association of FSHD-causing 4q35 deletions with a specific distal genomic sequence (4qA allele), altered DNA methylation patterns on 4q35, as well as other direct and indirect evidence point to epigenetic mechanisms. As a consequence, partial deletion of D4Z4 results in a (local) chromatin change and ultimately results in the loss of appropriate control of gene expression. There is at present no effective treatment for FSHD. A better understanding of the underlying pathophysiology is needed to design targeted interventions. Despite these limitations, however, two randomized controlled clinical trials have been conducted on FSHD. These trials, along with a previous natural history study, have helped to better define outcome measures for future trials in FSHD as well as other dystrophies.

Facioscapulohumeral muscular dystrophy

http://www.jbc.org/content/264/24/13998.full.pdf

A human muscle adenine nucleotide translocator gene has four exons, is located on chromosome 4, and is differentially expressed.

The human ATP synthase β subunit gene: Sequence analysis, chromosome assignment, and differential expression

The molybdenum hydroxylases are a ubiquitous class of enzymes which contain molybdenum in association with a low molecular weight cofactor. Genetic evidence suggests that the Drosophila loci, ma--1, cin and lxd are concerned with this cofactor because mutants for any one of these loci simultaneously interrupt activity for two molybdenum hydroxylases, XDH and A0. A third enzyme activity, P0, is also absent in each of the three mutants but evidence classifying P0 as a molybdoenzyme has been lacking. This study utilizes the known tungsten sensitivity of molybdoenzymes to demonstrate directly that pyridoxal oxidase is also molybdoenzyme. The low molecular weight molybdenum cofactor is found to be severely reduced in extracts of the 1xd and cin mutants but ma--1 mutants have high levels of cofactor. A partially purified preparation of XDH crossreacting material from ma--1 was also shown to contain the molybdenum cofactor. These results, considered with data from other workers are taken to indicate that the functions of all three of the loci examined could be concerned with some aspect of cofactor biosynthesis.

Molybdenum hydroxylases in Drosophila. II. Molybdenum cofactor in xanthine dehydrogenase, aldehyde oxidase and pyridoxal oxidase.

SummaryA reliable method for visualizing the Drosophila enzyme pyridoxal oxidase in polyacrylamide gels is described. Antiserum to pyridoxal oxidase has been produced and used in quantitative immunoelectrophoresis to determine the relative amounts of pyridoxal oxidase cross reacting material (CRM) in several mutants including lpo, lxd, ma-l and cin. The lpo variant did not have CRM for PO, thus further supporting the idea that it represents a structural gene for pyridoxal oxidase in Drosophila. CRM for PO was found in ma-l and lxd indicating that their effects upon the enzyme are probably post-translational. No CRM for PO could be found in the cin mutants.

http://link.springer.com/content/pdf/10.1007/BF00425861.pdf

Cynthia K. Warner

Papers

A human muscle adenine nucleotide translocator gene has four exons, is located on chromosome 4, and is differentially expressed.

The human ATP synthase β subunit gene: Sequence analysis, chromosome assignment, and differential expression

Molybdenum hydroxylases in Drosophila. II. Molybdenum cofactor in xanthine dehydrogenase, aldehyde oxidase and pyridoxal oxidase.

Molybdenum hydroxylases in Drosophila: I. Preliminary studies of pyridoxal oxidase

A genetic variant of beta-glucuronidase in Drosophila melanogaster.