Validation and development of MTH1 inhibitors for treatment of cancer

BACKGROUND

Previously, we showed cancer cells rely on the MTH1 protein to prevent incorporation of otherwise deadly oxidised nucleotides into DNA and we developed MTH1 inhibitors which selectively kill cancer cells. Recently, several new and potent inhibitors of MTH1 were demonstrated to be non-toxic to cancer cells, challenging the utility of MTH1 inhibition as a target for cancer treatment.

MATERIAL AND METHODS

Human cancer cell lines were exposed in vitro to MTH1 inhibitors or depleted of MTH1 by siRNA or shRNA. 8-oxodG was measured by immunostaining and modified comet assay. Thermal Proteome profiling, proteomics, cellular thermal shift assays, kinase and CEREP panel were used for target engagement, mode of action and selectivity investigations of MTH1 inhibitors. Effect of MTH1 inhibition on tumour growth was explored in BRAF V600E-mutated malignant melanoma patient derived xenograft and human colon cancer SW480 and HCT116 xenograft models.

RESULTS

Here, we demonstrate that recently described MTH1 inhibitors, which fail to kill cancer cells, also fail to introduce the toxic oxidized nucleotides into DNA. We also describe a new MTH1 inhibitor TH1579, (Karonudib), an analogue of TH588, which is a potent, selective MTH1 inhibitor with good oral availability and demonstrates excellent pharmacokinetic and anti-cancer properties in vivo.

CONCLUSION

We demonstrate that in order to kill cancer cells MTH1 inhibitors must also introduce oxidized nucleotides into DNA. Furthermore, we describe TH1579 as a best-in-class MTH1 inhibitor, which we expect to be useful in order to further validate the MTH1 inhibitor concept.

Classification and nomenclature of the superfamily of short-chain dehydrogenases/reductases (SDRs)

The short-chain dehydrogenases/reductases (SDRs) constitute one of the largest protein superfamilies known today. The members are distantly related with typically 20-30% residue identity in pair-wise comparisons. Still, all hitherto structurally known SDRs present a common three-dimensional structure consisting of a Rossmann fold with a parallel beta sheet flanked by three helices on each side. Using hidden Markov models (HMMs), we have developed a semi-automated subclassification system for this huge family. Currently, 75% of all SDR forms have been assigned to one of the 464 families totalling 122,940 proteins. There are 47 human SDR families, corresponding to 75 genes. Most human SDR families (35 families) have only one gene, while 12 have between 2 and 8 genes. For more than half of the human SDR families, the three-dimensional fold is known. The number of SDR members increases considerably every year, but the number of SDR families now starts to converge. The classification method has paved the ground for a sustainable and expandable nomenclature system. Information on the SDR superfamily is continuously updated at http://sdr-enzymes.org/.

The plant short-chain dehydrogenase (SDR) superfamily: genome-wide inventory and diversification patterns

BACKGROUND

Short-chain dehydrogenases/reductases (SDRs) form one of the largest and oldest NAD(P)(H) dependent oxidoreductase families. Despite a conserved 'Rossmann-fold' structure, members of the SDR superfamily exhibit low sequence similarities, which constituted a bottleneck in terms of identification. Recent classification methods, relying on hidden-Markov models (HMMs), improved identification and enabled the construction of a nomenclature. However, functional annotations of plant SDRs remain scarce.

RESULTS

Wide-scale analyses were performed on ten plant genomes. The combination of hidden Markov model (HMM) based analyses and similarity searches led to the construction of an exhaustive inventory of plant SDR. With 68 to 315 members found in each analysed genome, the inventory confirmed the over-representation of SDRs in plants compared to animals, fungi and prokaryotes. The plant SDRs were first classified into three major types - 'classical', 'extended' and 'divergent' - but a minority (10% of the predicted SDRs) could not be classified into these general types ('unknown' or 'atypical' types). In a second step, we could categorize the vast majority of land plant SDRs into a set of 49 families. Out of these 49 families, 35 appeared early during evolution since they are commonly found through all the Green Lineage. Yet, some SDR families - tropinone reductase-like proteins (SDR65C), 'ABA2-like'-NAD dehydrogenase (SDR110C), 'salutaridine/menthone-reductase-like' proteins (SDR114C), 'dihydroflavonol 4-reductase'-like proteins (SDR108E) and 'isoflavone-reductase-like' (SDR460A) proteins - have undergone significant functional diversification within vascular plants since they diverged from Bryophytes. Interestingly, these diversified families are either involved in the secondary metabolism routes (terpenoids, alkaloids, phenolics) or participate in developmental processes (hormone biosynthesis or catabolism, flower development), in opposition to SDR families involved in primary metabolism which are poorly diversified.

CONCLUSION

The application of HMMs to plant genomes enabled us to identify 49 families that encompass all Angiosperms ('higher plants') SDRs, each family being sufficiently conserved to enable simpler analyses based only on overall sequence similarity. The multiplicity of SDRs in plant kingdom is mainly explained by the diversification of large families involved in different secondary metabolism pathways, suggesting that the chemical diversification that accompanied the emergence of vascular plants acted as a driving force for SDR evolution.

Evolutionary conservation of the ribosomal biogenesis factor Rbm19/Mrd1: implications for function

Ribosome biogenesis in eukaryotes requires coordinated folding and assembly of a pre-rRNA into sequential pre-rRNA-protein complexes in which chemical modifications and RNA cleavages occur. These processes require many small nucleolar RNAs (snoRNAs) and proteins. Rbm19/Mrd1 is one such protein that is built from multiple RNA-binding domains (RBDs). We find that Rbm19/Mrd1 with five RBDs is present in all branches of the eukaryotic phylogenetic tree, except in animals and Choanoflagellates, that instead have a version with six RBDs and Microsporidia which have a minimal Rbm19/Mrd1 protein with four RBDs. Rbm19/Mrd1 therefore evolved as a multi-RBD protein very early in eukaryotes. The linkers between the RBDs have conserved properties; they are disordered, except for linker 3, and position the RBDs at conserved relative distances from each other. All but one of the RBDs have conserved properties for RNA-binding and each RBD has a specific consensus sequence and a conserved position in the protein, suggesting a functionally important modular design. The patterns of evolutionary conservation provide information for experimental analyses of the function of Rbm19/Mrd1. In vivo mutational analysis confirmed that a highly conserved loop 5-β4-strand in RBD6 is essential for function.

RSpred, a set of Hidden Markov Models to detect and classify the RIFIN and STEVOR proteins of Plasmodium falciparum

Background

Many parasites use multicopy protein families to avoid their host's immune system through a strategy called antigenic variation. RIFIN and STEVOR proteins are variable surface antigens uniquely found in the malaria parasites Plasmodium falciparum and P. reichenowi. Although these two protein families are different, they have more similarity to each other than to any other proteins described to date. As a result, they have been grouped together in one Pfam domain. However, a recent study has described the sub-division of the RIFIN protein family into several functionally distinct groups. These sub-groups require phylogenetic analysis to sort out, which is not practical for large-scale projects, such as the sequencing of patient isolates and meta-genomic analysis.

Results

We have manually curated the rif and stevor gene repertoires of two Plasmodium falciparum genomes, isolates DD2 and HB3. We have identified 25% of mis-annotated and ~30 missing rif and stevor genes. Using these data sets, as well as sequences from the well curated reference genome (isolate 3D7) and field isolate data from Uniprot, we have developed a tool named RSpred. The tool, based on a set of hidden Markov models and an evaluation program, automatically identifies STEVOR and RIFIN sequences as well as the sub-groups: A-RIFIN, B-RIFIN, B1-RIFIN and B2-RIFIN. In addition to these groups, we distinguish a small subset of STEVOR proteins that we named STEVOR-like, as they either differ remarkably from typical STEVOR proteins or are too fragmented to reach a high enough score. When compared to Pfam and TIGRFAMs, RSpred proves to be a more robust and more sensitive method. We have applied RSpred to the proteomes of several P. falciparum strains, P. reichenowi, P. vivax, P. knowlesi and the rodent malaria species. All groups were found in the P. falciparum strains, and also in the P. reichenowi parasite, whereas none were predicted in the other species.

Conclusions

We have generated a tool for the sorting of RIFIN and STEVOR proteins, large antigenic variant protein groups, into homogeneous sub-families. Assigning functions to such protein families requires their subdivision into meaningful groups such as we have shown for the RIFIN protein family. RSpred removes the need for complicated and time consuming phylogenetic analysis methods. It will benefit both research groups sequencing whole genomes as well as others working with field isolates. RSpred is freely accessible via http://www.ifm.liu.se/bioinfo/.

The SDR (short-chain dehydrogenase/reductase and related enzymes) nomenclature initiative

Short-chain dehydrogenases/reductases (SDR) constitute one of the largest enzyme superfamilies with presently over 46,000 members. In phylogenetic comparisons, members of this superfamily show early divergence where the majority have only low pairwise sequence identity, although sharing common structural properties. The SDR enzymes are present in virtually all genomes investigated, and in humans over 70 SDR genes have been identified. In humans, these enzymes are involved in the metabolism of a large variety of compounds, including steroid hormones, prostaglandins, retinoids, lipids and xenobiotics. It is now clear that SDRs represent one of the oldest protein families and contribute to essential functions and interactions of all forms of life. As this field continues to grow rapidly, a systematic nomenclature is essential for future annotation and reference purposes. A functional subdivision of the SDR superfamily into at least 200 SDR families based upon hidden Markov models forms a suitable foundation for such a nomenclature system, which we present in this paper using human SDRs as examples.

Prediction of coenzyme specificity in dehydrogenases/ reductases. A hidden Markov model-based method and its application on complete genomes

Dehydrogenases and reductases are enzymes of fundamental metabolic importance that often adopt a specific structure known as the Rossmann fold. This fold, consisting of a six-stranded beta-sheet surrounded by alpha-helices, is responsible for coenzyme binding. We have developed a method to identify Rossmann folds and predict their coenzyme specificity (NAD, NADP or FAD) using only the amino acid sequence as input. The method is based upon hidden Markov models and sequence pattern analysis. The prediction sensitivity is 79% and the selectivity close to 100%. The method was applied on a set of 68 genomes, representing the three kingdoms archaea, bacteria and eukaryota. In prokaryotes, 3% of the genes were found to code for Rossmann-fold proteins, while the corresponding ratio in eukaryotes is only around 1%. In all genomes, NAD is the most preferred cofactor (41-49%), followed by NADP with 30-38%, while FAD is the least preferred cofactor (21%). However, the NAD preponderance over NADP is most pronounced in archaea, and least in eukaryotes. In all three kingdoms, only 3-8% of the Rossmann proteins are predicted to have more than one membrane-spanning segment, which is much lower than the frequency of membrane proteins in general. Analysis of the major protein types in eukaryotes reveals that the most common type (26%) of the Rossmann proteins are short-chain dehydrogenases/reductases. In addition, the identified Rossmann proteins were analyzed with respect to further protein types, enzyme classes and redundancy. The described method is available at http://www.ifm.liu.se/bioinfo, where the preferred coenzyme and its binding region are predicted given an amino acid sequence as input.

Stabilization of discordant helices in amyloid fibril-forming proteins

Several proteins and peptides that can convert from alpha-helical to beta-sheet conformation and form amyloid fibrils, including the amyloid beta-peptide (Abeta) and the prion protein, contain a discordant alpha-helix that is composed of residues that strongly favor beta-strand formation. In their native states, 37 of 38 discordant helices are now found to interact with other protein segments or with lipid membranes, but Abeta apparently lacks such interactions. The helical propensity of the Abeta discordant region (K16LVFFAED23) is increased by introducing V18A/F19A/F20A replacements, and this is associated with reduced fibril formation. Addition of the tripeptide KAD or phospho-L-serine likewise increases the alpha-helical content of Abeta(12-28) and reduces aggregation and fibril formation of Abeta(1-40), Abeta(12-28), Abeta(12-24), and Abeta(14-23). In contrast, tripeptides with all-neutral, all-acidic or all-basic side chains, as well as phosphoethanolamine, phosphocholine, and phosphoglycerol have no significant effects on Abeta secondary structure or fibril formation. These data suggest that in free Abeta, the discordant alpha-helix lacks stabilizing interactions (likely as a consequence of proteolytic removal from a membrane-associated precursor protein) and that stabilization of this helix can reduce fibril formation.

Coenzyme-based functional assignments of short-chain dehydrogenases/reductases (SDRs)

Short-chain dehydrogenases/reductases (SDRs) are enzymes of great functional diversity. In spite of a residue identity of only 15-30%, the folds are conserved to a large extent, with specific sequence motifs detectable. We have developed an assignment scheme based on these motifs and detect five families. Only two of these were known before, called 'Classical' and 'Extended', but are now distinguished at a further level based on patterns of charged residues in the coenzyme-binding region, giving seven subfamilies of classical SDRs and three subfamilies of extended SDRs. Three further families are novel entities, denoted 'Intermediate', 'Divergent' and 'Complex', encompassing short-chain alcohol dehydrogenases, enoyl reductases and multifunctional enzymes, respectively. The assignment scheme was applied to the genomes of human, mouse, D. melanogaster, C. elegans, A. thaliana and S. cerevisiae. In the animal genomes, genes corresponding to the extended SDRs amount to around one quarter or less of the total number of SDR genes, while in those of A. thaliana and S. cerevisiae, the extended members constitute about 40% of the SDR forms. The NAD(H)-dependent SDRs are about equally many as the NADP(H)-dependent ones in human, mouse and plant, while the proportions of NAD(H)-dependent enzymes are much lower in fruit fly, worm and yeast. We also find that NADP(H) is the preferred coenzyme among most classical SDRs, while NAD(H) is that preferred among most extended SDRs.

Short-chain dehydrogenases/reductases (SDR): the 2002 update

Short-chain dehydrogenases/reductases (SDR) form a large, functionally heterogeneous protein family presently with about 3000 primary and about 30 3D structures deposited in databases. Despite low sequence identities between different forms (about 15-30%), the 3D structures display highly similar alpha/beta folding patterns with a central beta-sheet, typical of the Rossmann-fold. Based on distinct sequence motifs functional assignments and classifications are possible, making it possible to build a general nomenclature system. Recent mutagenetic and structural studies considerably extend the knowledge on the general reaction mechanism, thereby establishing a catalytic tetrad of Asn-Ser-Tyr-Lys residues, which presumably form the framework for a proton relay system including the 2'-OH of the nicotinamide ribose, similar to the mechanism found in horse liver ADH. Based on their cellular functions, several SDR enzymes appear as possible and promising pharmacological targets with application areas spanning hormone-dependent cancer forms or metabolic diseases such as obesity and diabetes, and infectious diseases.

Short-chain dehydrogenases/reductases (SDRs)

Short-chain dehydrogenases/reductases (SDRs) are enzymes of great functional diversity. Even at sequence identities of typically only 15-30%, specific sequence motifs are detectable, reflecting common folding patterns. We have developed a functional assignment scheme based on these motifs and we find five families. Two of these families were known previously and are called 'classical' and 'extended' families, but they are now distinguished at a further level based on coenzyme specificities. This analysis gives seven subfamilies of classical SDRs and three subfamilies of extended SDRs. We find that NADP(H) is the preferred coenzyme among most classical SDRs, while NAD(H) is that preferred among most extended SDRs. Three families are novel entities, denoted 'intermediate', 'divergent' and 'complex', encompassing short-chain alcohol dehydrogenases, enoyl reductases and multifunctional enzymes, respectively. The assignment scheme was applied to the genomes of human, mouse, Drosophila melanogaster, Caenorhabditis elegans, Arabidopsis thaliana and Saccharomyces cerevisiae. In the animal genomes, the extended SDRs amount to around one quarter or less of the total number of SDRs, while in the A. thaliana and S. cerevisiae genomes, the extended members constitute about 40% of the SDR forms. The numbers of NAD(H)-dependent and NADP(H)-dependent SDRs are similar in human, mouse and plant, while the proportions of NAD(H)-dependent enzymes are much lower in fruit fly, worm and yeast. We show that, in spite of the great diversity of the SDR superfamily, the primary structure alone can be used for functional assignments and for predictions of coenzyme preference.

Short-chain dehydrogenase/reductase (SDR) relationships: a large family with eight clusters common to human, animal, and plant genomes

The progress in genome characterizations has opened new routes for studying enzyme families. The availability of the human genome enabled us to delineate the large family of short-chain dehydrogenase/reductase (SDR) members. Although the human genome releases are not yet final, we have already found 63 members. We have also compared these SDR forms with those of three model organisms: Caenorhabditis elegans, Drosophila melanogaster, and Arabidopsis thaliana. We detect eight SDR ortholog clusters in a cross-genome comparison. Four of these clusters represent extended SDR forms, a subgroup found in all life forms. The other four are classical SDRs with activities involved in cellular differentiation and signalling. We also find 18 SDR genes that are present only in the human genome of the four genomes studied, reflecting enzyme forms specific to mammals. Close to half of these gene products represent steroid dehydrogenases, emphasizing the regulatory importance of these enzymes.

Prediction of amyloid fibril-forming proteins

In Alzheimer's disease and spongiform encephalopathies proteins transform from their native states into fibrils. We find that several amyloid-forming proteins harbor an alpha-helix in a polypeptide segment that should form a beta-strand according to secondary structure predictions. In 1324 nonredundant protein structures, 37 beta-strands with > or =7 residues were predicted in segments where the experimentally determined structures show helices. These discordances include the prion protein (helix 2, positions 179-191), the Alzheimer amyloid beta-peptide (Abeta, positions 16-23), and lung surfactant protein C (SP-C, positions 12-27). In addition, human coagulation factor XIII (positions 258-266), triacylglycerol lipase from Candida antarctica (positions 256-266), and d-alanyl-d-alanine transpeptidase from Streptomyces R61 (positions 92-106) contain a discordant helix. These proteins have not been reported to form fibrils but in this study were found to form fibrils in buffered saline at pH 7.4. By replacing valines in the discordant helical part of SP-C with leucines, an alpha-helix is found experimentally and by secondary structure predictions. This analogue does not form fibrils under conditions where SP-C forms abundant fibrils. Likewise, when Abeta residues 14-23 are removed or changed to a nondiscordant sequence, fibrils are no longer formed. We propose that alpha-helix/beta-strand-discordant stretches are associated with amyloid fibril formation.

KIND-a non-redundant protein database

SUMMARY

KIND (Karolinska Institutet Nonredundant Database) is a protein database where identical sequences, both full length and partial, have been removed. The database contains nearly 274 900 sequences, half of which originate from the protein sequence databases Swissprot and PIR, while the other half come from translated open reading frames in GenPept and TrEMBL.

AVAILABILITY

KIND is downloadable from ftp://ftp.mbb.ki.se/pub/KIND.