A Computational Framework for Proteome-Wide Pursuit and Prediction of Metalloproteins using ICP-MS and MS/MS Data
- W Andrew Lancaster†1,
- Jeremy L Praissman†1,
- Farris L PooleII1,
- Aleksandar Cvetkovic1,
- Angeli Lal Menon1,
- Joseph W Scott1,
- Francis E JenneyJr1, 2,
- Michael P Thorgersen1,
- Ewa Kalisiak3,
- Junefredo V Apon3,
- Sunia A Trauger3,
- Gary Siuzdak3,
- John A Tainer4 and
- Michael W W Adams1Email author
© Lancaster et al; licensee BioMed Central Ltd. 2011
Received: 10 October 2010
Accepted: 28 February 2011
Published: 28 February 2011
Metal-containing proteins comprise a diverse and sizable category within the proteomes of organisms, ranging from proteins that use metals to catalyze reactions to proteins in which metals play key structural roles. Unfortunately, reliably predicting that a protein will contain a specific metal from its amino acid sequence is not currently possible. We recently developed a generally-applicable experimental technique for finding metalloproteins on a genome-wide scale. Applying this metal-directed protein purification approach (ICP-MS and MS/MS based) to the prototypical microbe Pyrococcus furiosus conclusively demonstrated the extent and diversity of the uncharacterized portion of microbial metalloproteomes since a majority of the observed metal peaks could not be assigned to known or predicted metalloproteins. However, even using this technique, it is not technically feasible to purify to homogeneity all metalloproteins in an organism. In order to address these limitations and complement the metal-directed protein purification, we developed a computational infrastructure and statistical methodology to aid in the pursuit and identification of novel metalloproteins.
We demonstrate that our methodology enables predictions of metal-protein interactions using an experimental data set derived from a chromatography fractionation experiment in which 870 proteins and 10 metals were measured over 2,589 fractions. For each of the 10 metals, cobalt, iron, manganese, molybdenum, nickel, lead, tungsten, uranium, vanadium, and zinc, clusters of proteins frequently occurring in metal peaks (of a specific metal) within the fractionation space were defined. This resulted in predictions that there are from 5 undiscovered vanadium- to 13 undiscovered cobalt-containing proteins in Pyrococcus furiosus. Molybdenum and nickel were chosen for additional assessment producing lists of genes predicted to encode metalloproteins or metalloprotein subunits, 22 for nickel including seven from known nickel-proteins, and 20 for molybdenum including two from known molybdo-proteins. The uncharacterized proteins are prime candidates for metal-based purification or recombinant approaches to validate these predictions.
We conclude that the largely uncharacterized extent of native metalloproteomes can be revealed through analysis of the co-occurrence of metals and proteins across a fractionation space. This can significantly impact our understanding of metallobiochemistry, disease mechanisms, and metal toxicity, with implications for bioremediation, medicine and other fields.
Metallomics is an emerging field that seeks to comprehensively characterize the role of metals in organisms . As with any new field, unique challenges have been encountered in terms of experimental methodologies and data analysis. The essential role metals play in biology has long been appreciated, but the complete metallome of any organism has yet to be characterized. It is estimated that around a third of all proteins in an organism require a metal partner . While it is possible to predict that certain proteins contain a metal , there are fundamental limitations to all current computational methods in comprehensively describing the metalloproteome of any organism. Many current methods rely on sequence motifs that in turn depend on the existence of a sufficiently sized set of previously annotated homologous proteins, a problem further compounded by the diversity of metal-binding sites across organisms as well as within a single organism . It is also well known that heterologous protein expression can result in the production of proteins incorporating metals that are not natively incorporated . In addition, while certain proteins and protein families are known to bind a variety of metals and are annotated accordingly, many are annotated as binding a single metal based on limited evidence.
It was shown recently  that the set of metals known to interact with proteins in vivo is a significant underestimate of the true extent and diversity of the metalloproteome. The approach developed used metal-directed protein purification relying on inductively coupled plasma mass spectrometry (ICP-MS) and tandem mass spectrometry (MS/MS) and revealed that a prototypical microbe, Pyrococcus furiosus, takes up 21 of 53 metals measured in its growth medium, 18 of which are present in macromolecular complexes. These results are in stark contrast to the five metals that had previously been identified in proteins individually purified from the same organism. Further, of the 343 metal peaks found across the fractions from a second level of chromatography fractionation (for the 10 metals that were detected), almost half (158) contained no known or predicted metalloprotein corresponding to that particular metal . The purification of eight of these metal peaks resulted in the identification of novel metalloproteins, or proteins containing unexpected metal ions . Unfortunately, this method has two major limitations. Firstly, given the large number (158) of unassigned metal peaks and difficulty in purifying a single protein, it is impractical to purify a significant portion of these novel metalloproteins. Secondly, it is not technically feasible to natively purify proteins of very low abundance over several chromatographic steps.
Herein is described a computational infrastructure and analytical methodology developed to both aid in the pursuit of novel metalloproteins  as well as to predict which proteins observed via MS/MS during this fractionation are likely to be metalloproteins without requiring purification to homogeneity. This required the development of a database, an Online Analytical Processing (OLAP) cube and InterPro-Metal (IPM) automated metal domain identification methods (directly supporting the pursuit of novel metalloproteins), as well as a Global Metal Protein Association (GMPA) analysis (enabling the prediction of metal-protein associations without complete purification). Given the essential biological role of metals, the discovery of novel metalloproteins has a multitude of implications in a variety of fields [1, 8]. Moreover, the computational infrastructure and methods described can be applied to any form of biomass of interest from tissues to microbes to identify potential metalloprotein targets for experimental characterization. Most importantly, this analysis allows the discovery of low abundance metalloproteins without radioisotope labeling, which have eluded other methods [9, 10] but which nonetheless may occupy key roles in essential biological pathways.
The experimental data set utilized in this study is an expanded version of the data set described in . Briefly, native biomass of the hyperthermophilic archaeon Pyrococcus furiosus was fractionated anaerobically through multiple non-denaturing chromatography steps utilizing multiple column chemistries. The resulting 2,589 fractions were analyzed by ICP-MS to identify metals, and by MS/MS to identify proteins (primarily high-throughput MS/MS in this study). The MS/MS data were filtered such that the false discovery rate was less than 1%, as described in  and only proteins identified by Mascot with two or more peptides were considered in the current bioinformatics based study. The use of non-denaturing native chromatography, ICP-MS and MS/MS captures the co-occurrence of metals and proteins in their native form, and enabled a metal-based purification strategy, in contrast to conventional enzyme assay guided protein purification . While this metal-based separation was used to purify a number of metalloproteins to homogeneity , the wealth of metal and protein data collected for proteins that were not explicitly targeted for direct purification provided an additional opportunity (applying data analysis techniques) to identify proteins that are likely to contain one or more metals in their native form.
A relational database was constructed using Microsoft SQL Server 2005 to store the data used in this study. The database consists of three principal modules: a procedural (fractionation) module, a metal data (ICP-MS) module, and a protein data (MS/MS) module. The fractionation module was designed to store the procedural information used in each of the separation steps carried out during the multi-level column fractionation (allowing reconstruction of the complete experimental pathway). These multi-level hierarchical relationships between fractions were queried using recursive common table expressions (CTE). The metal data module was designed to store both procedural data and replicate metal concentration data for each sample and metal analyzed using ICPMS. This module also stores the peak assignments determined by manual inspection. Finally, the protein module stores data for each peptide identified by MS/MS, its protein source and the corresponding ORF, and details related to the MS/MS run and Mascot search as imported from Mascot XML result files. All fractions and samples (from fractions) collected were assigned unique IDs and labeled with 2 D data matrix barcodes to facilitate sample tracking. This ensured the simple and reliable association of the data contained in all three principal modules within the context of the experimental hierarchy.
Bioinformatic metalloprotein prediction-InterPro-Metal (IPM) automated metal domain analysis
The set of known metalloproteins that have been previously purified and characterized from P. furiosus by conventional chromatographic methods consists of only 23 proteins (encoded by 39 genes). Each contains one or more of Co, Fe, Ni, W and/or Zn atoms . Although the utility of bioinformatic predictions is limited, such predictions can be used to identify homologs of more extensively characterized metalloproteins and serve as a starting point for assigning proteins to the observed metal peaks. The Integrated Resource of Protein Domains and Functional Sites (InterPro)  was used to predict known metal associated domains encoded in the genome of P. furiosus. InterPro integrates multiple popular protein feature databases, and provides the Iprscan utility for searching protein sequence queries against these databases. The genome was searched using this utility and the resulting matches of proteins to InterPro entries were stored in a relational database (protein-InterPro data). The description of each InterPro entry, including name, abstract and publication list, is available in a downloadable XML file ftp://ftp.ebi.ac.uk/pub/databases/interpro/interpro.xml.gz. This file was parsed and inserted into a corresponding relational database. A number of regular expression patterns relating to metal ions, metal cofactors and metal binding domains were used to search the text of each InterPro entry description to classify the entries which potentially involve specific metals (metal-InterPro data). Those metal-InterPro entries that had hits in the P. furiosus genome were evaluated manually for quality and assigned a subjective score. In some cases, keyword hits were not deemed to be indicative of a potential association of the given domain with a given metal, for instance an abstract for a particular subfamily of an enzyme may include additional information on other subfamilies which use alternate metals. Such spurious hits were assigned a score of 0 while metals with evidence of association with the given domain were assigned a positive score. All hits with nonzero scores were considered as potentially metal associated domains in subsequent analyses. The protein-InterPro data and metal-InterPro data were joined to determine which P. furiosus proteins had associations with specific metals and this subsequently will be referred to as the InterPro-Metal (IPM) database or InterPro-Metal analysis. These domain-based predictions were incorporated into the relational database and OLAP cube (Figure 1) to aid in the identification of novel metalloproteins and proteins for which the metal prediction and observed metal associations differ.
Data driven metalloprotein prediction-Global Metal Protein Association (GMPA) analysis
Results and Discussion
The computational framework that was developed consisted of a database, an Online Analytical Processing (OLAP) cube, InterPro-Metal (IPM) automated metal domain identification and Global Metal Protein Association (GMPA) analysis. This complemented and enhanced our recent effort to elucidate the metalloproteome of P. furiosus and to identify novel metalloproteins . The GMPA analysis in particular was used to provide estimates of the numbers of metalloproteins that could be expected proteome-wide and a narrowed list of candidates (based on ICP-MS and MS/MS data) at various stages of the column chromatography fractionation, culminating in our predictions at the conclusion of the study that the numbers of undiscovered metal containing proteins in P. furiosus range from approximately 5 for vanadium up to as many as 13 for cobalt. Validation of these predictions is provided by the overall success of the GMPA analysis in categorizing known metalloproteins from P. furiosus, the establishment of lower bounds on the numbers of proteins required to explain all metal peaks seen during the fractionation, the fact that the predictions lie within reasonable ranges in the context of literature (where up to a third of proteins are expected to contain or be involved with metals in various ways with the caveat that the majority are likely to be involved with Mg) [2, 17, 18] and considering the effect of dynamic association/adventitious metal binding .
In order to determine how much of the information contained in our data set remains uncaptured by the GMPA analysis, nickel (Ni) and molybdenum (Mo) were chosen for manual evaluation using the GMPA predictions as a starting point. Of the 870 proteins identified by MS/MS with two or more peptides, 153 and 119 were found to be significantly associated with Ni and Mo respectively, upon clustering yielding predictions of 13 and 10 total Ni- and Mo-proteins in the proteome of P. furiosus. The local semi-quantitative MS/MS peptide hit data for each of the proteins clustered was then manually compared to the local metal concentration data using our data explorer (Figure 1). This step excluded an additional 131 and 99 proteins producing top candidates lists of 22 and 20 proteins that are most likely to contain Ni and Mo respectively. These lists were then analyzed more extensively through bioinformatic analyses and literature searches. We will first describe the results obtained at the conclusion of our experimental study and then discuss the bioinformatics of the lists of predicted Ni- and Mo-proteins, concluding with the limitations inherent in this study.
Bioinformatic metalloprotein prediction-InterPro-Metal (IPM) automated metal domain analysis predictions
Data driven metalloprotein prediction-Global Metal Protein Association (GMPA) analysis predictions
GMPA Analysis Clustering and Coverage of Known Metalloproteins
Known metalloprotein subunits
Nickel- and Molybdo-protein evaluation
Manually Evaluated Nickel Protein Candidates
Aldolase-type TIM barrel
Alba archaeal DNA/RNA-binding protein
Beta-lactamase-like glyoxalase II family member
Glycosyl transferase, family 2
Conserved hypothetical protein
Carbohydrate binding protein
Uncharacterized rubrerythrin domain protein
Hydrogenase II beta
Hydrogenase II delta
Hydrogenase II alpha
Hydrogenase I beta
Hydrogenase I alpha
Pyroxidine biosynthesis protein
Peptidyl-prolyl cis-trans isomerase
Alanyl-tRNA synthetase, class IIc
Hydrogenase expression/formation protein A
Manually Evaluated Molybdenum Protein Candidates
ThiF family protein
Putative cofactor synthesis protein
Wyosine base formation, Radical SAM
Cell division transporter FtsY
Fructose-1,6-bisphosphate aldolase class I
Protein of unknown function DUF1621
Peptidyl-prolyl cis-trans isomerase
Signal recognition particle 54
Protein of unknown function DUF509
Protein of unknown function DUF217
Protein of unknown function DUF89
DNA polymerase, family B
Translation factor, SUA5 type
Bioinformatic analyses of predicted Ni- and Mo-proteins
As discussed above, 7 of the 22 genes listed in Table 2 encode proteins or subunits of proteins which have been shown to contain Ni ions in P. furiosus. This includes subunits of soluble hydrogenase I (SHI) and soluble hydrogenase 2 (SHII) grouped in clusters 7 and 8 respectively. In addition, PF0615 in cluster 13 is annotated as a hypA protein, which is implicated in Ni insertion in the hydrogenases. The structure of a hypA homolog from Thermococcus kodakaraensis has been solved and its Ni-binding site described . This demonstrates that cluster 13 has at least two Ni-binding proteins that frequently co-occur in the fractionation space. Of the five ORFs in Table 2 with homologs whose crystal structures have been solved bound to metals other than Ni, three are now known to bind Ni (PF0056 and PF0086) in P. furiosus or are known to have a Ni-binding site (PF0615). In particular PF0086 has been shown to bind Ni , but its homolog from the closely related Pyrococcus horikoshii (PDB 2E18) was expressed recombinantly and crystallized in a Zn-bound form. This illustrates the flexibility of metal binding domains , and their ability to bind biologically incorrect metals when expressed heterologously . The two remaining ORFs with non-Ni homolog crystal structures are PF1664, which contains the cysteines that bind Zn (Cd in the crystal) in its homolog  and may be involved in binding Ni in P. furiosus, and PF2038 with a homolog that binds Mg-ATP. The only protein listed in Table 2 that is likely not to contain Ni is PF1861. This was previously purified from P. furiosus biomass and contained Co and Zn but not Ni . This leaves 13 proteins that are predicted to contain Ni. These proteins have no known or conjecturable Ni associations based on their sequences and are assumed to predominantly contain a set of undiscovered Ni-binding sites. Finally, it is worth pointing out that PF0056 is one of five ORFs (PF0144, PF1987, PF0056, PF0138 and PF1500) annotated as either conserved hypothetical or with only domain/motif matches and has now been shown to bind Ni .
In contrast to the case of Ni, the pool of known molybdo-proteins in P. furiosus is small and far less can be ascertained bioinformatically. In particular, the role of Mo-proteins in P. furiosus is unclear, with only two such metalloproteins having recently been identified . Consequently, only two of 20 proteins in Table 3 have either been shown to bind Mo (PF1587) or have an IPM hit for Mo (PF0187). The other recently identified Mo-protein that was purified from P. furiosus (PF1972) was observed in only 17 chromatography fractions and was rejected by the GMPA analysis, which depends on sufficient levels of occurrence in the data set to establish significance of metal-protein association. On the positive side, many of the uncharacterized proteins contain residues that could be involved in Mo-binding (e.g. 14 of 20 contain at least one cysteine residue as is often involved in Mo-pterin binding) [26–28], but given the extent and complexity of typical molybdopterin binding interactions and biochemistry  and the lack of knowledge of Mo-binding in organisms closely related to P. furiosus, we have not looked into this aspect further. On the negative side, DNA polymerase (Mo cluster 9, Table 3), which has been well studied in many different organisms and is not known to bind or utilize Mo (although it is not clear if this has been directly considered previously) was picked as a top candidate Mo-protein. This illustrates that some of the targets that appear to reliably co-occur with a metal may be coincidental, or the result of interaction natively with additional proteins that are not strictly required for their primary function. Interestingly, four of the 20 predicted Mo-protein candidates have annotations that include "domain of unknown function." The confirmation of Mo-binding, which has already occurred for PF1587 by metal-directed purification, should provide an improved foundation for functionally characterizing these conserved domains which so far have been elusive .
We initially attempted to use standard correlation-based statistical techniques such as principal component analysis (PCA) and canonical correlation analysis (CCA) to determine associations between metals and proteins based on the experimental data that were available . However, these efforts were hindered by the relatively non-quantitative nature of the MS/MS data available (lacking even spectral count information). Consequently, the GMPA analysis method was developed which is less reliant on quantitative agreement between these data sets. Simulated data sets demonstrated the effectiveness of GMPA scores alone given adequate separation regardless of the amount of noise observed in the peptide counts, but it was discovered that the metal-based fractionation typically did not produce a comprehensive enough data set containing an appropriate degree of overall separation. For example, the experimental data set is most consistent and comprehensive at the second column level (termed C2 in ) and separation is still relatively incomplete at this level. Consequently significance cut-off curves and clustering were employed completing the overall GMPA analysis methodology. It is likely that the predictive power of the methodology could be greatly improved by utilizing a data set with a more comprehensive fractionation across all levels, through the use of more quantitative MS/MS techniques [30–32] and more powerful statistical techniques (PCA/CCA) that could then be applied more easily. This methodology could also potentially be carried out in a more automated fashion on an analytical scale to provide a rapid determination of the metalloproteins of any organism.
We have presented a computational methodology that can uncover probable metal-containing proteins using data from a non-comprehensive native fractionation coupled with metal and protein measurement using ICP-MS and MS/MS. This methodology has identified a number of candidate novel metalloproteins that are targets for future experimental verification. Application of the method to simulated data sets indicates that additional predictive accuracy could be achieved through the use of a more comprehensive fractionation. Our results for each of the 10 metals examined in this study underscore the unexplored complexity of metalloproteomes and have broad implications for protein structure and function as well as metal toxicity.
high-throughput tandem mass spectrometry
inductively coupled plasma mass spectrometry
tandem mass spectrometry
online analytical processing
open reading frame
This work conducted by ENIGMA was supported by the Office of Science, Office of Biological and Environmental Research, of the U. S. Department of Energy under Contract No. DE-AC02-05CH11231.
- Mounicou S, Szpunar J, Lobinski R: Metallomics: the concept and methodology. Chemical Society Reviews 2009, 38: 1119–1138. 10.1039/b713633cView ArticlePubMedGoogle Scholar
- Waldron KJ, Rutherford JC, Ford D, Robinson NJ: Metalloproteins and metal sensing. Nature 2009, 460: 823–830. 10.1038/nature08300View ArticlePubMedGoogle Scholar
- Andreini C, Bertini I, Rosato A: Metalloproteomes: A Bioinformatic Approach. Acc Chem Res 2009, 42: 1471–1479. 10.1021/ar900015xView ArticlePubMedGoogle Scholar
- Dupont CL, Butcher A, Valas RE, Bourne PE, Caetano-Anolles G: History of biological metal utilization inferred through phylogenomic analysis of protein structures. Proc Natl Acad Sci USA 2010, 107: 10567–10572. 10.1073/pnas.0912491107PubMed CentralView ArticlePubMedGoogle Scholar
- Waldron KJ, Robinson NJ: How do bacterial cells ensure that metalloproteins get the correct metal? Nat Rev Microbiol 2009, 7: 25–35. 10.1038/nrmicro2057View ArticlePubMedGoogle Scholar
- Cvetkovic A, Menon AL, Thorgersen MP, Scott JW, Poole FL, Jenney FE Jr, Lancaster WA, Praissman JL, Shanmukh S, Vaccaro BJ, et al.: Microbial metalloproteomes are largely uncharacterized. Nature 2010, 466: 779–782. 10.1038/nature09265View ArticlePubMedGoogle Scholar
- Menon AL, Poole FL, Cvetkovic A, Trauger SA, Kalisiak E, Scott JW, Shanmukh S, Praissman J, Jenney FE Jr, Wikoff WR, et al.: Novel multiprotein complexes identified in the hyperthermophilic archaeon Pyrococcus furiosus by non-denaturing fractionation of the native proteome. Mol Cell Proteomics 2009, 8: 735–751. 10.1074/mcp.M800246-MCP200PubMed CentralView ArticlePubMedGoogle Scholar
- Haferburg G, Kothe E: Metallomics: lessons for metalliferous soil remediation. Appl Microbiol Biotechnol 2010, 87: 1271–1280. 10.1007/s00253-010-2695-zView ArticlePubMedGoogle Scholar
- Tottey S, Waldron KJ, Firbank SJ, Reale B, Bessant C, Sato K, Cheek TR, Gray J, Banfield MJ, Dennison C, Robinson NJ: Protein-folding location can regulate manganese-binding versus copper- or zinc-binding. Nature 2008, 455: 1138–1142. 10.1038/nature07340View ArticlePubMedGoogle Scholar
- Sevcenco AM, Krijger GC, Pinkse MW, Verhaert PD, Hagen WR, Hagedoorn PL: Development of a generic approach to native metalloproteomics: application to the quantitative identification of soluble copper proteins in Escherichia coli . J Biol Inorg Chem 2009, 14: 631–640. 10.1007/s00775-009-0477-9View ArticlePubMedGoogle Scholar
- Kehl C, Simms AM, Toofanny RD, Daggett V: Dynameomics: a multi-dimensional analysis-optimized database for dynamic protein data. Protein Eng Des Sel 2008, 21: 379–386. 10.1093/protein/gzn015View ArticlePubMedGoogle Scholar
- Hunter S, Apweiler R, Attwood TK, Bairoch A, Bateman A, Binns D, Bork P, Das U, Daugherty L, Duquenne L, et al.: InterPro: the integrative protein signature database. Nucleic Acids Res 2009, 37: D211–215. 10.1093/nar/gkn785PubMed CentralView ArticlePubMedGoogle Scholar
- R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900051–07–0[http://www.R-project.org]
- Langfelder P, Zhang B, Horvath S: Defining clusters from a hierarchical cluster tree: the Dynamic Tree Cut package for R. Bioinformatics 2008, 24: 719–720. 10.1093/bioinformatics/btm563View ArticlePubMedGoogle Scholar
- Zhang B, Horvath S: A general framework for weighted gene co-expression network analysis. Stat Appl Genet Mol Biol 2005, 4: Article17.PubMedGoogle Scholar
- Chvatal V: A Greedy Heuristic for the Set-Covering Problem. Mathematics of operations research 1979, 4: 233–235. 10.1287/moor.4.3.233View ArticleGoogle Scholar
- Andreini C, Bertini I, Cavallaro G, Holliday GL, Thornton JM: Metal ions in biological catalysis: from enzyme databases to general principles. Journal Of Biological Inorganic Chemistry 2008, 13: 1205–1218. 10.1007/s00775-008-0404-5View ArticlePubMedGoogle Scholar
- Yamaguchi A, Iida K, Matsui N, Tomoda S, Yura K, Go M: Het-PDB Navi.: a database for protein-small molecule interactions. J Biochem 2004, 135: 79–84. 10.1093/jb/mvh009View ArticlePubMedGoogle Scholar
- Xiao Z, Wedd AG: The challenges of determining metal-protein affinities. Nat Prod Rep 2010, 27: 768–789. 10.1039/b906690jView ArticlePubMedGoogle Scholar
- Pruitt KD, Tatusova T, Maglott DR: NCBI reference sequences (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res 2007, 35: D61–65. 10.1093/nar/gkl842PubMed CentralView ArticlePubMedGoogle Scholar
- Watanabe S, Arai T, Matsumi R, Atomi H, Imanaka T, Miki K: Crystal structure of HypA, a nickel-binding metallochaperone for [NiFe] hydrogenase maturation. J Mol Biol 2009, 394: 448–459. 10.1016/j.jmb.2009.09.030View ArticlePubMedGoogle Scholar
- Dudev T, Lim C: Metal Binding Affinity and Selectivity in Metalloproteins: Insights from Computational Studies. Annual Review of Biophysics 2008, 37: 97–116. 10.1146/annurev.biophys.37.032807.125811View ArticlePubMedGoogle Scholar
- Jenney FE Jr, Adams MW: Rubredoxin from Pyrococcus furiosus . Methods Enzymol 2001, 334: 45–55. full_textView ArticlePubMedGoogle Scholar
- Sivaraman J, Myers RS, Boju L, Sulea T, Cygler M, Jo Davisson V, Schrag JD: Crystal structure of Methanobacterium thermoautotrophicum phosphoribosyl-AMP cyclohydrolase HisI. Biochemistry 2005, 44: 10071–10080. 10.1021/bi050472wView ArticlePubMedGoogle Scholar
- Story SV, Shah C, Jenney FE Jr, Adams MW: Characterization of a novel zinc-containing, lysine-specific aminopeptidase from the hyperthermophilic archaeon Pyrococcus furiosus . J Bacteriol 2005, 187: 2077–2083. 10.1128/JB.187.6.2077-2083.2005PubMed CentralView ArticlePubMedGoogle Scholar
- Brondino CD, Romao MJ, Moura I, Moura JJ: Molybdenum and tungsten enzymes: the xanthine oxidase family. Curr Opin Chem Biol 2006, 10: 109–114. 10.1016/j.cbpa.2006.01.034View ArticlePubMedGoogle Scholar
- Gladyshev VN, Khangulov SV, Axley MJ, Stadtman TC: Coordination of selenium to molybdenum in formate dehydrogenase H from Escherichia coli . Proc Natl Acad Sci USA 1994, 91: 7708–7711. 10.1073/pnas.91.16.7708PubMed CentralView ArticlePubMedGoogle Scholar
- Romao MJ, Knablein J, Huber R, Moura JJ: Structure and function of molybdopterin containing enzymes. Prog Biophys Mol Biol 1997, 68: 121–144. 10.1016/S0079-6107(97)00022-9View ArticlePubMedGoogle Scholar
- Schwarz G, Mendel RR, Ribbe MW: Molybdenum cofactors, enzymes and pathways. Nature 2009, 460: 839–847. 10.1038/nature08302View ArticlePubMedGoogle Scholar
- Lu P, Vogel C, Wang R, Yao X, Marcotte EM: Absolute protein expression profiling estimates the relative contributions of transcriptional and translational regulation. Nat Biotechnol 2007, 25: 117–124. 10.1038/nbt1270View ArticlePubMedGoogle Scholar
- Silva JC, Gorenstein MV, Li GZ, Vissers JP, Geromanos SJ: Absolute quantification of proteins by LCMSE: a virtue of parallel MS acquisition. Mol Cell Proteomics 2006, 5: 144–156.View ArticlePubMedGoogle Scholar
- Zhu W, Smith JW, Huang CM: Mass spectrometry-based label-free quantitative proteomics. J Biomed Biotechnol 2010, 2010: 840518.PubMed CentralPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.