MBF1 (multiprotein bridging factor 1) is a highly conserved protein in archaea and eukaryotes. It was originally identified as a mediator of the eukaryotic transcription regulator BmFTZ-F1 (Bombyx mori regulator of fushi tarazu). MBF1 was demonstrated to enhance transcription by forming a bridge between distinct regulatory DNA-binding proteins and the TATA-box-binding protein. MBF1 consists of two parts: a C-terminal part that contains a highly conserved helix–turn–helix, and an N-terminal part that shows a clear divergence: in eukaryotes, it is a weakly conserved flexible domain, whereas, in archaea, it is a conserved zinc-ribbon domain. Although its function in archaea remains elusive, its function as a transcriptional co-activator has been deduced from thorough studies of several eukaryotic proteins, often indicating a role in stress response. In addition, MBF1 was found to influence translation fidelity in yeast. Genome context analysis of mbf1 in archaea revealed conserved clustering in the crenarchaeal branch together with genes generally involved in gene expression. It points to a role of MBF1 in transcription and/or translation. Experimental data are required to allow comparison of the archaeal MBF1 with its eukaryotic counterpart.
Because most information-processing pathways in eukaryotes are highly similar to that of archaea, it is tempting to use the generally less complex archaeal systems as models. The archaeal transcription machinery, for example, is highly similar to the eukaryotic RNAP (RNA polymerase) II machinery. The core subunits of the RNA polymerase are present in all archaea, together with orthologues of some general transcription factors: TBP (TATA-box-binding protein), TF (transcription factor) IIB (TFB in archaea), and the α-subunit of TFIIE (TFE in archaea) . In contrast, the majority of the known eukaryotic transcription factors are absent. A protein that is reported to be present in almost all archaea and known as a transcriptional co-activator in eukaryotes is MBF1 (multiprotein bridging factor 1) . The function of this protein in archaea remains poorly understood. However, its function as a transcriptional co-activator in eukaryotes has been described in numerous studies since it was discovered in Bombyx mori . Interestingly, MBF1 has also been reported to influence fidelity of translation in yeast . In the present paper, we review the current knowledge of MBF1 and its possible role(s) in archaea: is MBF1 involved in transcription or in translation, or does it have a dual function?
MBF1 was first reported in 1994 in B. mori as a mediator of BmFTZ-F1, the equivalent of the FTZ-F1 regulator of the Drosophila fushi tarazu (ftz) gene that is associated with body segmentation. It was demonstrated that a heterodimer of MBF1 and MBF2 (which is not related to MBF1 and is found only in moths), could enhance transcription of a FTZ-F1-dependent transcript by bridging between BmFTZ-F1 and TBP [3,5–8]. Early genomic comparisons revealed that a gene coding for MBF1 is present in all eukaryotic genomes [2,6,9–11]. In humans (h), MBF1 is present in two isoforms: hMBF1α and hMBF1β, which originates from alternative splicing. The α mRNA was found in all the tissues tested, whereas the β form was only detected in pancreas and HeLa cells. The β form contains a very acidic tail, possibly an adaptation to the pancreatic environment . The plant Arabidopsis thaliana (At) has three distinct copies in its genome, encoding AtMBF1a, AtMBF1b and AtMBF1c. All three paralogues appear to have distinct biological roles . These distinct MBF1 proteins and MBF1 proteins of other plants can be divided into two subgroups on the basis of their primary structure and different exon/intron patterns .
As to the distribution of MBF1 among prokaryotes, none of the sequenced bacterial genomes contains an orthologue of mbf1, whereas almost all archaea are equipped with one. This seems a logical distribution for a transcriptional co-activator: eukaryotes and the archaea have a high similarity in their transcription initiation complex, whereas the bacterial system is very different. Most archaeal genomes contain a single copy of the mbf1 gene; however, all of the sequenced genomes of the Halobacteriales contain a second copy that lacks the C-terminal amino acids (Figure 1).
Schematic overview of conserved domains in eukaryotic and archaeal MBF1
Interestingly, the archaeal genomes that do not possess an mbf1 gene are the only two species that are completely sequenced within the recently proposed new phylum of the Thaumarchaeota : Cenarchaeum symbiosum and Nitrosopumilus maritimus. Apparently, the gene is lost in this archaeal branch (Figure 2).
Genome context analysis of mbf1 in archaea
Sequence and structure
In a comparative genomics analysis of HTH (helix–turn–helix) domains, which are generally responsible for protein–DNA interactions, Aravind and Koonin  found that most of the predicted HTH-transcriptional regulators found in archaea appeared to be of the bacterial type. This came as a surprise because of the aforementioned relationship between the archaeal transcription initiation system and the eukaryotic RNAPII. It was reported that the only conserved classical HTH domain that is vertically inherited in almost all archaea and all eukaryotes was MBF1. NMR studies of hMBF1 and BmMBF1 indeed revealed a well structured C-terminal HTH core. It contains four α-helices and a conserved tail, proposed to be responsible for maintaining domain stability [16,17]. Binding of MBF1 to DNA-binding regulators occurs via the central region, residues 35–113 in B. mori, that also contains the HTH core . Binding of MBF1 to TBP in yeast (y) requires an aspartate residue at position 112 (Asp112) of yMBF1, located in the third helix of the C-terminal part, and a glutamine residue at position 68 (Gln68) of yTBP, located on top of the saddle-shaped molecule. In eukaryotic and archaeal species, several combinations of amino acids are present on the corresponding locations. Altered yMBF1 and yTBP proteins with combinations that exist in Nature show an in vitro interaction, whereas unnatural combinations (e.g. Gln68/Lys112 and Glu68/Asp112) do not, suggesting co-evolution of MBF1 and TBP . This holds true for the archaeal combinations, although experimental evidence for an interaction between archaeal TBP and MBF1 is currently lacking.
In rats, MBF1 interacts with calmodulin . This is in agreement with the presence in mammalian MBF1 of an IQ motif that interacts with calmodulin in the cytoplasm [19,20]. Binding is increased by calcium and appears to be a function that is distinct from its involvement in transcription activation, as impairing one does not impair the other . Both calmodulin and the IQ motif are absent from archaea. A phosphorylation site is located in close proximity to the IQ motif. Phosphorylation of MBF1 disrupts calmodulin binding and leads to translocation to the nucleus [19,20]. In mammals, phosphorylation also appears to be necessary for MBF1-potentiated expression in vivo during cardiac hypertrophy, which is a stress response to increased workload . Phosphorylation has also been reported for potato MBF1. It could be increased by elicitors derived from Phytophthora infestans and is inhibited by calcium chelators and calmodulin antagonists, suggesting that phosphorylation of MBF1 is carried out by a calcium-dependent serine/threonine kinase . Unlike its mammalian counterpart, plant MBF1 appears to be localized predominantly in the nucleus .
In contrast with the high homology found in the C-terminal part of MBF1, the N-terminal part shows a clear divergence between eukaryotes and archaea. In eukaryotes, it has low conservation, and NMR studies indicate a relatively high flexibility [16,17]. In archaea, it contains a conserved zinc-ribbon motif, predicted on the basis of its two pairs of cysteine residues. The presence of this zinc ribbon in archaea and its absence in eukaryotes could indicate distinct functions. Whereas, in eukaryotes, MBF1 binding to DNA by itself was not observed, the presence of a zinc ribbon might indicate that archaeal MBF1 functions as a single activator in archaea . However, at present, no experimental support for this proposal is available.
After its first description, MBF1 has been studied in several eukaryotic organisms. It does not affect histone modification, but rather acts as a bridging factor between a DNA-binding protein and the entire TFIID complex [TBP and TAFs (TBP-associated factors)] [3,5,6,12,17,19,20,24–27]. DNA-binding proteins that have been reported to interact with MBF1 belong to two protein families: the family of steroid/nuclear hormone receptors, and the family of the bZIPs (basic leucine zippers). Established interactions between MBF1 and members of the former family are FTZ-F1, Ad4BP (adrenal 4-binding protein; a human orthologue of FTZ-F1), and liver receptor homologue 1, liver X receptor α and PPARγ (peroxisome-proliferator-activated receptor γ), non-steroid nuclear receptors, involved in lipid metabolism in humans [3,5–7,25]. Known interaction partners belonging to the bZIP family include GCN4 (general control non-derepressible 4), CREB (cAMP-response-element-binding protein) and CREB1, ATF (activating transcription factor) 1, Jun and Fos, TDF (tracheae defective factor) and HaHB4 (Helianthus annuus homeobox-leucine zipper protein 4) [12,21,24,25,28,29]. Jun, Fos and ATF family members dimerize into homo- or hetero-dimers, all called AP-1 (activator protein 1). Different AP-1 dimers have different physiological effects. It has been reported, for example, that, for activation of the atrial natriuretic peptide, MBF1 interacts with c-Jun, but not with c-Fos or ATF, suggesting that MBF1 may be able to discriminate between highly similar proteins. Interestingly, MBF1 is not always essential, as other promoters can be activated by c-Jun in the absence of MBF1 [21,30].
In yeast, MBF1 is essential for GCN4-dependent activation of HIS3, encoding an imidazoleglycerol-phosphate dehydratase that is responsible for a step in histidine biosynthesis. Without activation, a basic level of constitutive expression of this gene still allows growth. However disruption of mbf1 generates a phenotype that is sensitive to 3-aminotriazole, an inhibitor of HIS3 enzyme . In human endothelial cell cultures, MBF1 represses endothelial cell differentiation. Therefore hMBF1 is also called EDF-1 (endothelial cell differentiation factor 1) [10,31]. In Drosophila, mbf1-null mutants are viable, but show severe tracheal and central nervous system defects. However, no segmentation defects similar to those of ftz-f1 mutants and no decrease in expression of a FTZ-F1-dependent reporter gene were observed, indicating that MBF1 is not a crucial co-activator for FTZ-F1-dependent transcription in vivo in fruitflies . These reports indicate involvement of MBF1 in different physiological processes in phylogenetically distinct eukaryotic organisms. Although it is completely conserved in eukaryotes, mbf1 knockouts appear viable under conditions tested.
Unexpectedly, mbf1 was found to be synonymous with the frameshift-suppression gene suf13 in yeast . The suf13-1 mutant strains as well as a directed suf13 (mbf1)-gene-knockout strain exhibit an increased rate of ribosomal +1 frameshifting for several different reporter gene constructs harbouring frameshift mutations [4,32,33]. It remains elusive how a transcription co-activator influences translation fidelity. The authors speculate about a role of MBF1 in RNA polymerase III transcription of tRNA genes . Decreased levels of tRNA would increase translational pausing and thereby the probability of frameshift events. TBP is also part of the RNAPIII general transcription factor TFIIIB. Similarly to its role in RNAPII transcription activation, MBF1 could therefore interact with the RNAPIII transcription machinery via TBP, but, up to now, no experimental evidence links MBF1 to RNAPIII transcription initiation.
An alternative scenario would be an involvement of MBF1 in the biogenesis of ribosomes and tRNA, possibly indirectly by transcription regulation of factors contributing to these pathways. In addition, it cannot be ruled out that MBF1 interacts directly with the ribosome during translation as it is also present in the cytoplasm [23,28,34,35]. Both translation and tRNA gene transcription are central parts of information processing. A function of MBF1 in these processes would explain the remarkable evolutionary conservation of MBF1 in eukaryotes and archaea.
in vivo analyses in plants indicate a link between MBF1 and stress response, including the ethylene response. MBF1s from Arabidopsis, potato, tomato, Retama raetam (white weeping broom) and tobacco have been reported to be up-regulated upon biotic and abiotic stresses, such as pathogen infection, drought, salinity, application of ethylene and, especially, heat and oxidative stress. AtMBF1c also reacts on application of abscisic acid and salicylic acid. It controls the heat-stress-response network. AtMBF1a is linked to salt stress, and is also, together with AtMBF1b, differentially regulated during plant development [11,14,22,23,27,35–40]. Overexpression in plants leads to a higher tolerance to a number of stress factors, without suppressing plant growth. It also leads to higher levels of trehalose, which is generally believed to be important in stress signalling in plants [35,40,41]. In Drosophila, it reduces sensitivity to oxidative stress by protecting the transcriptional activator d-Jun against oxidation . The latter effect is found in mammals too, where MBF1 appears to be a crucial factor in the response against oxidative stress . During hyperthrophy, enhanced mbf1 gene expression leads to elevated protein levels in mammalian cardiomyocyte cultures, and, in turn, results in activation of hypertrophic gene expression . Interestingly, MBF1 appears also to be a target for viral abuse: MBF1 of tobacco (Nicotiana tabacum) (NtMBF1a) and Arabidopsis (AtMBF1a and AtMBF1b) have been reported to interact with two viral movement proteins, responsible for cell-to-cell movement .
In eukaryotes, MBF1 plays an important role in the stress response. This might also be true for the archaeal MBF1, although mbf1 is not up-regulated during oxidative stress (B. Wiedenheft, M. Young and J. van der Oost, unpublished work).
In prokaryotes, genomic context analysis is a powerful tool in the functional prediction of genes . Especially in crenarchaeotes, the gene context of mbf1 shows a high degree of conservation encompassing several genes (with predicted function): pan (proteasome-activating nucleotidase ), hflX (G-protein of the HflX family ), tgt (tRNA-guanine transglycosylase ), tfe and tfb (coding for archaeal homologues of transcription initiation factors E and B respectively) and rpoG (hypothetical RNA-polymerase subunit G ) (Figure 2). The proteasome-activating nucleotidase is homologous with the ATPases of the 19S particle of the eukaryotic proteasome and regulates the entry of folded proteins into the proteolytic core particle of the proteasome. The HflX family of G-proteins is largely uncharacterized, but a function linked to the translation machinery has been predicted [48,49]. In euryarchaeotes, the genomic context of mbf1 is less conserved; however, proximity to pan occurs regularly. Apart from this, the genomic context is only conserved at higher taxonomic levels. Interestingly, although mbf1 is absent from Thaumarchaeota, a part of the chromosome shows similarities to the genomic context of mbf1 in Thermofilum pendens, except that mbf1 has been replaced by a gene encoding a putative phosphate-uptake regulator.
It is not possible to give a clear functional prediction based on the observed context in crenarchaeotes. However, the conserved clustering (and probably the co-regulation) with genes involved in gene expression at different levels points to a basal function of MBF1 in gene expression as well, probably at the level of translation and/or transcription.
Functional prediction for archaeal MBF1
MBF1 is a conserved protein present in all eukaryotes and archaea, except the two characterized members of the Thaumarchaeota. The protein consists of an N-terminal part that differs between eukaryotes and archaea, and a very well conserved C-terminal part. The observation that the N-terminal domain of human MBF1α is not required for binding to TBP  suggests that this interaction can also be found in archaea. Such a role would fit with the absence of MBF1 in the bacterial domain where the transcription machinery is very different, and with the aforementioned indications of MBF1/TBP co-evolution . However the divergence in the N-terminal part may point to a different mode of action, possibly a direct interaction with DNA. This is supported by the absence of bZIP proteins in the archaeal domain. An interesting other possibility is the finding that yMBF1 (directly or indirectly) influences translation fidelity.
The genomic context is certainly interesting, but unfortunately not conclusive either. The strongest co-occurrence of archaeal mbf1 is with pan, the gene encoding a protein associated with protein degradation. Other conserved genes, mainly found in crenarchaeotes, encode proteins in one way or another associated with transcription or translation. Potential experimental approaches to gain insight into the physiological function of the archaeal MBF1 include a broad spectrum of complementary biochemical, genetic and genomic studies (in vivo and in vitro): comparison of wild-type, mbf1 knockout and MBF1 overproduction become strained at the level of transcription and translation. In the absence of such experimental data, however, comparison of the archaeal MBF1 with its eukaryotic counterpart still seems to be a bridge too far.
This work was financially supported by a Vici grant from the Dutch Organization for Scientific Research (Nederlandse Organisatie voor Wetenschappelijk Onderzoek) [grant number 865.05.001].
Molecular Biology of Archaea: Biochemical Society Focused Meeting held at University of St Andrews, U.K., 19–21 August 2008. Organized and Edited by Stephen Bell (Oxford, U.K.) and Malcolm White (St Andrews, U.K.).
activator protein 1
activating transcription factor
basic leucine zipper
regulator of fushi tarazu
general control non-derepressible 4
multiprotein bridging factor 1
Present address: Protein Studies Program, Oklahoma Medical Research Foundation, 825 Northeast Thirteenth Street, Oklahoma City, OK 73104, U.S.A.