The viral coat protein is regulated by HSP70 and HSP40 in Potato virus A infection

Plus-stranded (plus) RNA viruses multiply within a cellular environment as tightly integrated units and rely on the genetic information carried within their genomes for multiplication and, hence, persistence. The minimal genomes of plus RNA viruses are unable to encode the molecular machineries that are required for virus multiplication. This sets requisites for the virus, which must form compatible interactions with host components during multiplication to successfully utilize primary metabolites as building blocks or metabolic energy, and to divert the protein synthesis machinery for production of viral proteins. In fact, the emerging picture of a virus-infected cell displays tight integration with the virus, from simple host and virus protein interactions through to major changes in the physiological state of the host cell. This study set out to develop a method for the identification of host components, mainly host proteins, that interact with proteins of Potato virus A (PVA; Potyvirus) during infection. This goal was approached by developing affinity-tag based methods for the purification of viral proteins complexed with associated host proteins from infected plants. Using this method, host membrane-associated viral ribonucleoprotein (RNP) complexes were obtained, and several host and viral proteins could be identified as components of these complexes. One of the host proteins identified using this strategy was a member of the heat shock protein 70 (HSP70) family, and this protein was chosen for further analysis. To enable the analysis of viral gene expression, a second method was developed based on Agrobacterium-mediated virus genome delivery into plant cells, and detection of virally expressed Renilla luciferase (RLUC) as a quantitative measure of viral gene expression. Using this method, it was observed that down-regulation of HSP70 caused a PVA coat protein (CP)-mediated defect associated with replication. Further experimentation suggested that CP can inhibit viral gene expression and that a distinct translational activity coupled to replication, referred to as replication-associated translation (RAT), exists. Unlike translation of replication-deficient viral RNA, RAT was dependent on HSP70 and its co-chaperone CPIP. HSP70 and CPIP together regulated CP turnover by promoting its modification by ubiquitin. Based on these results, an HSP70 and CPIP-driven mechanism that functions to regulate CP during viral RNA replication and/or translation is proposed, possibly to prevent premature particle assembly caused by CP association with viral RNA.

Limiting the Extent of the RDN1 Heterochromatin Domain by a Silencing Barrier and Sir2 Protein Levels in Saccharomyces cerevisiae

In Saccharomyces cerevisiae, transcriptional silencing occurs at the cryptic mating-type loci (HML and HMR), telomeres, and ribosomal DNA ( rDNA; RDN1). Silencing in the rDNA is unusual in that polymerase II (Pol II) promoters within RDN1 are repressed by Sir2 but not Sir3 or Sir4. rDNA silencing unidirectionally spreads leftward, but the mechanism of limiting its spreading is unclear. We searched for silencing barriers flanking the left end of RDN1 by using an established assay for detecting barriers to HMR silencing. Unexpectedly, the unique sequence immediately adjacent to RDN1, which overlaps a prominent cohesin binding site (CARL2), did not have appreciable barrier activity. Instead, a fragment located 2.4 kb to the left, containing a tRNA(Gln) gene and the Ty1 long terminal repeat, had robust barrier activity. The barrier activity was dependent on Pol III transcription of tRNA(Gln), the cohesin protein Smc1, and the SAS1 and Gcn5 histone acetyltransferases. The location of the barrier correlates with the detectable limit of rDNA silencing when SIR2 is overexpressed, where it blocks the spreading of rDNA heterochromatin. We propose a model in which normal Sir2 activity results in termination of silencing near the physical rDNA boundary, while tRNA(Gln) blocks silencing from spreading too far when nucleolar Sir2 pools become elevated.

Characterization of LRRTM and NGR gene families: Expression and functions

Some leucine-rich repeat (LRR) -containing membrane proteins are known regulators of neuronal growth and synapse formation. In this work I characterize two gene families encoding neuronal LRR membrane proteins, namely the LRRTM (leucine-rich repeat, transmembrane neuronal) and NGR (Nogo-66 receptor) families. I studied LRRTM and NGR family member's mRNA tissue distribution by RT-PCR and by in situ hybridization. Subcellular localization of LRRTM1 protein was studied in neurons and in non-neuronal cells. I discovered that LRRTM and NGR family mRNAs are predominantly expressed in the nervous system, and that each gene possesses a specific expression pattern. I also established that LRRTM and NGR family mRNAs are expressed by neurons, and not by glial cells. Within neurons, LRRTM1 protein is not transported to the plasma membrane; rather it localizes to endoplasmic reticulum. Nogo-A (RTN4), MAG, and OMgp are myelin-associated proteins that bind to NgR1 to limit axonal regeneration after central nervous system injury. To better understand the functions of NgR2 and NgR3, and to explore the possible redundancy in the signaling of myelin inhibitors of neurite growth, I mapped the interactions between NgR family and the known and candidate NgR1 ligands. I identified high-affinity interactions between RTN2-66, RTN3-66 and NgR1. I also demonstrate that Rtn3 mRNA is expressed in the same glial cell population of mouse spinal cord white matter as Nogo-A mRNA, and thus it could have a role in myelin inhibition of axonal growth. To understand how NgR1 interacts with multiple structurally divergent ligands, I aimed first to map in more detail the nature of Nogo-A:NgR1 interactions, and then to systematically map the binding sites of multiple myelin ligands in NgR1 by using a library of NgR1 expression constructs encoding proteins with one or multiple surface residues mutated to alanine. My analysis of the Nogo-A:NgR1 -interactions revealed a novel interaction site between the proteins, suggesting a trivalent Nogo-A:NgR1-interaction. Our analysis also defined a central binding region on the concave side of NgR1's LRR domain that is required for the binding of all known ligands, and a surrounding region critical for binding MAG and OMgp. To better understand the biological role of LRRTMs, I generated Lrrtm1 and Lrrtm3 knock out mice. I show here that reporter genes expressed from the targeted loci can be used for maping the neuronal connections of Lrrtm1 and Lrrtm3 expressing neurons in finer detail. With regard to LRRTM1's role in humans, we found a strong association between a 70 kb-spanning haplotype in the proposed promoter region of LRRTM1 gene and two possibly related phenotypes: left-handedness and schizophrenia. Interestingly, the responsible haplotype was linked to phenotypic variability only when paternally inherited. In summary, I identified two families of neuronal receptor-like proteins, and mapped their expression and certain protein-protein interactions. The identification of a central binding region in NgR1 shared by multiple ligands may facilitate the design and development of small molecule therapeutics blocking binding of all NgR1 ligands. Additionally, the genetic association data suggests that allelic variation upstream of LRRTM1 may play a role in the development of left-right brain asymmetry in humans. Lrrtm1 and Lrrtm3 knock out mice developed as a part of this study will likely be useful for schizophrenia and Alzheimer s disease research.

How Effective Is The Data On Co-Occurrence Of Domains In Multi-Domain Proteins In Prediction Of Protein-Protein Interactions?

We explore the fuse of information on co-occurrence of domains in multi-domain proteins in predicting protein-protein interactions. The basic premise of our work is the assumption that domains co-occurring in a polypeptide chain undergo either structural or functional interactions among themselves. In this study we use a template dataset of domains in multidomain proteins and predict protein-protein interactions in a target organism. We note that maximum number of correct predictions of interacting protein domain families (158) is made in S. cerevisiae when the dataset of closely related organisms is used as the template followed by the more diverse dataset of bacterial proteins (48) and a dataset of randomly chosen proteins (23). We conclude that use of multi-domain information from organisms closely-related to the target can aid prediction of interacting protein families.

Heat-shock protein 70-2 (HSP70-2) expression in bladder urothelial carcinoma is associated with tumour progression and promotes migration and invasion

Purpose: Testis specific heat-shock protein 70-2 (HSP70-2), a member of HSP70 chaperone family, is essential for the growth of spermatocytes and cancer cells. We investigated the association of HSP70-2 expression with clinical behaviour and progression of urothelial carcinoma of bladder. Experimental design: We assessed the HSP70-2 expression by RT-PCR and HSP70-2 protein expression by immunofluorescence, flow cytometry, immunohistochemistry and Western blotting in urothelial carcinoma patient specimens and HTB-1, UMUC-3, HTB-9, HTB-2 and normal human urothelial cell lines. Further, to investigate the role of HSP70-2 in bladder tumour development, HSP70-2 was silenced in the high-grade invasive HTB-1 and UMUC-3 cells. The malignant properties of urothelial carcinoma cells were examined using colony formation, migration assay, invasion assay in vitro and tumour growth in vivo. Results: Our RT-PCR analysis and immunohistochemistry analysis revealed that HSP70-2 was expressed in both moderate to well-differentiated and high-grade invasive urothelial carcinoma cell lines studied and not in normal human urothelial cells. In consistence with these results, HSP70-2 expression was also observed in superficially invasive (70%) and muscle-invasive (90%) patient's tumours. Furthermore, HSP70-2 knockdown significantly suppressed cellular motility and invasion ability. An in vivo xenograft study showed that inhibition of HSP70-2 significantly suppressed tumour growth. Conclusions: In conclusion, our data suggest that the HSP70-2 expression is associated with early spread and progression of urothelial carcinoma of bladder cancer and that HSP70-2 can be the potential therapeutic target for bladder urothelial carcinoma. (C) 2009 Elsevier Ltd. All rights reserved.

The Structure and Functions of Hantavirus Nucleocapsid Protein

Hantaviruses, members of the genus Hantavirus in the Bunyaviridae family, are enveloped single-stranded RNA viruses with tri-segmented genome of negative polarity. In humans, hantaviruses cause two diseases, hemorrhagic fever with renal syndrome (HFRS) and hantavirus pulmonary syndrome (HPS), which vary in severity depending on the causative agent. Each hantavirus is carried by a specific rodent host and is transmitted to humans through excreta of infected rodents. The genome of hantaviruses encodes four structural proteins: the nucleocapsid protein (N), the glycoproteins (Gn and Gc), and the polymerase (L) and also the nonstructural protein (NSs). This thesis deals with the functional characterization of hantavirus N protein with regard to its structure. Structural studies of the N protein have progressed slowly and the crystal structure of the whole protein is still not available, therefore biochemical assays coupled with bioinformatical modeling proved essential for studying N protein structure and functions. Presumably, during RNA encapsidation, the N protein first forms intermediate trimers and then oligomers. First, we investigated the role of N-terminal domain in the N protein oligomerization. The results suggested that the N-terminal region of the N protein forms a coiled-coil, in which two antiparallel alpha helices interact via their hydrophobic seams. Hydrophobic residues L4, I11, L18, L25 and V32 in the first helix and L44, V51, L58 and L65 in the second helix were crucial for stabilizing the structure. The results were consistent with the head-to-head, tail-to-tail model for hantavirus N protein trimerization. We demonstrated that an intact coiled-coil structure of the N terminus is crucial for the oligomerization capacity of the N protein. We also added new details to the head-to-head, tail-to-tail model of trimerization by suggesting that the initial step is based on interaction(s) between intact intra-molecular coiled-coils of the monomers. We further analyzed the importance of charged aa residues located within the coiled-coil for the N protein oligomerization. To predict the interacting surfaces of the monomers we used an upgraded in silico model of the coiled-coil domain that was docked into a trimer. Next the predicted target residues were mutated. The results obtained using the mammalian two-hybrid assay suggested that conserved charged aa residues within the coiled-coil make a substantial contribution to the N protein oligomerization. This contribution probably involves the formation of interacting surfaces of the N monomers and also stabilization of the coiled-coil via intramolecular ionic bridging. We proposed that the tips of the coiled-coils are the first to come into direct contact and thus initiate tight packing of the three monomers into a compact structure. This was in agreement with the previous results showing that an increase in ionic strength abolished the interaction between N protein molecules. We also showed that residues having the strongest effect on the N protein oligomerization are not scattered randomly throughout the coiled-coil 3D model structure, but form clusters. Next we found evidence for the hantaviral N protein interaction with the cytoplasmic tail of the glycoprotein Gn. In order to study this interaction we used the GST pull-down assay in combination with mutagenesis technique. The results demonstrated that intact, properly folded zinc fingers of the Gn protein cytoplasmic tail as well as the middle domain of the N protein (that includes aa residues 80 248 and supposedly carries the RNA-binding domain) are essential for the interaction. Since hantaviruses do not have a matrix protein that mediates the packaging of the viral RNA in other negatve stranded viruses (NSRV), hantaviral RNPs should be involved in a direct interaction with the intraviral domains of the envelope-embedded glycoproteins. By showing the N-Gn interaction we provided the evidence for one of the crucial steps in the virus replication at which RNPs are directed to the site of the virus assembly. Finally we started analysis of the N protein RNA-binding region, which is supposedly located in the middle domain of the N protein molecule. We developed a model for the initial step of RNA-binding by the hantaviral N protein. We hypothesized that the hantaviral N protein possesses two secondary structure elements that initiate the RNA encapsidation. The results suggest that amino acid residues (172-176) presumably act as a hook to catch vRNA and that the positively charged interaction surface (aa residues 144-160) enhances the initial N-RNA interacation. In conclusion, we elucidated new functions of hantavirus N protein. Using in silico modeling we predicted the domain structure of the protein and using experimental techniques showed that each domain is responsible for executing certain function(s). We showed that intact N terminal coiled-coil domain is crucial for oligomerization and charged residues located on its surface form a interaction surface for the N monomers. The middle domain is essential for interaction with the cytoplasmic tail of the Gn protein and RNA binding.

Functions of Alphavirus Macrodomain-containing protein nsP3

Alphaviruses are positive strand RNA viruses that replicate in association with cellular membranes. The viral RNA replication complex consists of four non-structural proteins nsP1-nsP4 which are essential for viral replication. The functions of nsP1, nsP2 and nsP4 are well established, but the roles of nsP3 are mainly unknown. In this work I have clarified some of the functions of nsP3 in order to better understand the importance of this protein in virus replication. Semliki Forest virus (SFV) has been mostly used as a model alphavirus during this work, but some experiments have also been conducted with Sindbis and Chikungunya viruses. NsP3 is composed of three different protein domains. The N-terminus of nsP3 contains an evolutionarily conserved macrodomain, the central part of nsP3 contains a domain that is only found in alphaviruses, and the C-terminus of the protein is hypervariable and predicted to be unstructured. In this work I have analyzed the functions of nsP3 macrodomain, and shown that viral macrodomains bind poly(ADP-ribose) and that they do not resemble cellular macrodomains in their properties. Furthermore, I have shown that some macrodomains, including viral macrodomains of SFV and hepatitis E virus, also bind poly(A). Mutations in the ligand binding pocket of SFV macrodomain hamper virus replication but do not confer lethality, indicating that macrodomain function is beneficial but not mandatory for virus replication. The hypervariable C-terminus of nsP3 is heavily phosphorylated and is enriched in proline residues. In this work it is shown that this region harbors an SH3 domain binding motif (Sh3BM) PxRxPR through which cellular amphiphysin is recruited to viral replication sites and to nsP3 containing cytoplasmic aggregate structures. The function of Sh3BM was destroyed by a single point mutation, which led to impaired viral RNA replication in HeLa cells, pointing out the functional importance of amphiphysin recruitment by the Sh3BM. In addition, evidence is provided tho show that the endosomal localization of alphavirus replication is mediated by nsP3 and that the phosphorylation of hypervariable region might be important for the endosomal targeting. Together these findings demonstrate that nsP3 contains multiple important host interaction motifs and domains, which facilitate successful viral propagation in host cells.

The Critical Role of N- and C-Terminal Contact in Protein Stability and Folding of a Family 10 Xylanase under Extreme Conditions

Background: Stabilization strategies adopted by proteins under extreme conditions are very complex and involve various kinds of interactions. Recent studies have shown that a large proportion of proteins have their N- and C-terminal elements in close contact and suggested they play a role in protein folding and stability. However, the biological significance of this contact remains elusive. Methodology: In the present study, we investigate the role of N- and C-terminal residue interaction using a family 10 xylanase (BSX) with a TIM-barrel structure that shows stability under high temperature,alkali pH, and protease and SDS treatment. Based on crystal structure,an aromatic cluster was identified that involves Phe4, Trp6 and Tyr343 holding the Nand C-terminus together; this is a unique and important feature of this protein that might be crucial for folding and stabilityunder poly-extreme conditions. Conclusion: A series of mutants was created to disrupt this aromatic cluster formation and study the loss of stability and function under given conditions. While the deletions of Phe4 resulted in loss of stability, removal of Trp6 and Tyr343 affected in vivo folding and activity. Alanine substitution with Phe4, Trp6 and Tyr343 drastically decreased stability under all parameters studied. Importantly,substitution of Phe4 with Trp increased stability in SDS treatment.Mass spectrometry results of limited proteolysis further demonstrated that the Arg344 residue is highly susceptible to trypsin digestion in sensitive mutants such as DF4, W6A and Y343A, suggesting again that disruption of the Phe4-Trp6-Tyr343 (F-W-Y) cluster destabilizes the N-and C-terminal interaction. Our results underscore the importance of N- and C-terminal contact through aromatic interactions in protein folding and stability under extreme conditions, and these results may be useful to improve the stability of other proteins under suboptimal conditions.

Biochemical characterization of the intracellular domain of the human guanylyl cyclase C receptor provides evidence for a catalytically active homotrimer

Guanylyl cyclase C (GCC) is the receptor for the family of guanylin peptides and bacterial heat-stable enterotoxins (ST). The receptor is composed of an extracellular, ligand-binding domain and an intracellular domain with a region of homology to protein kinases and a guanylyl cyclase catalytic domain. We have expressed the entire intracellular domain of GCC in insect cells and purified the recombinant protein, GCC-IDbac, to study its catalytic activity and regulation. Kinetic properties of the purified protein were similar to that of full-length GCC, and high activity was observed when MnGTP was used as the substrate. Nonionic detergents, which stimulate the guanylyl cyclase activity of membrane-associated GCC, did not appreciably increase the activity of GCC-IDbac, indicating that activation of the receptor by Lubrol involved conformational changes that required the transmembrane and/or the extracellular domain. The guanylyl cyclase activity of GCC-IDbac was inhibited by Zn2+, at concentrations shown to inhibit adenylyl cyclase, suggesting a structural homology between the two enzymes. Covalent crosslinking of GCC-IDbac indicated that the protein could associate as a dimer, but a large fraction was present as a trimer. Gel filtration analysis also showed that the major fraction of the protein eluted at a molecular size of a trimer, suggesting that the dimer detected by cross-linking represented subtle differences in the juxtaposition of the individual polypeptide chains. We therefore provide evidence that the trimeric state of GCC is catalytically active, and sequences required to generate the trimer are present in the intracellular domain of GCC.

cAMP-regulated Protein Lysine Acetylases in Mycobacteria

Cyclic AMP synthesized by Mycobacterium tuberculosis has been shown to play a role in pathogenesis. However, the high levels of intracellularcAMP found in both pathogenic and nonpathogenic mycobacteria suggest that additional and important biological processes are regulated by characterization of novel cAMP-binding proteins in M. smegmatis and M. tuberculosis (MSMEG_5458 and Rv0998, respectively) that contain a cyclic nucleotide binding domain fused to a domain that shows similarity to the GNAT family of acetyltransferases. We detect protein lysine acetylation in mycobacteria and identify a universal stress protein (USP) as a substrate of MSMEG_5458. Acetylation of a lysine residue in USP is regulated by cAMP, and using a strain deleted for MSMEG_5458, we show that USP is indeed an in vivo substrate for MSMEG_5458. The Rv0998 protein shows a strict cAMP-dependent acetylation of USP, despite a lower affinity for cAMP than MSMEG_5458. Thus, this report not only represents the first demonstration of protein lysine acetylation in mycobacteria but also describes a unique functional interplay between a cyclic nucleotide binding domain and a protein acetyltransferase.

Genome-wide analysis and experimentation of plant serine/threonine/tyrosine-specific protein kinases

Protein tyrosine phosphorylation plays an important role in cell growth, development and oncogenesis. No classical protein tyrosine kinase has hitherto been cloned from plants. Does protein tyrosine kinase exist in plants? To address this, we have performed a genomic survey of protein tyrosine kinase motifs in plants using the delineated tyrosine phosphorylation motifs from the animal system. The Arabidopsis thaliana genome encodes 57 different protein kinases that have tyrosine kinase motifs. Animal non-receptor tyrosine kinases, SRC, ABL, LYN, FES, SEK, KIN and RAS have structural relationship with putative plant tyrosine kinases. In an extended analysis, animal receptor and non-receptor kinases, Raf and Ras kinases, mixed lineage kinases and plant serine/threonine/tyrosine (STY) protein kinases, form a well-supported group sharing a common origin within the superfamily of STY kinases. We report that plants lack bona fide tyrosine kinases, which raise an intriguing possibility that tyrosine phosphorylation is carried out by dual-specificity STY protein kinases in plants. The distribution pattern of STY protein kinase families on Arabidopsis chromosomes indicates that this gene family is partly a consequence of duplication and reshuffling of the Arabidopsis genome and of the generation of tandem repeats. Genome-wide analysis is supported by the functional expression and characterization of At2g24360 and phosphoproteomics of Arabidopsis. Evidence for tyrosine phosphorylated proteins is provided by alkaline hydrolysis, anti-phosphotyrosine immunoblotting, phosphoamino acid analysis and peptide mass fingerprinting. These results report the first comprehensive survey of genome-wide and tyrosine phosphoproteome analysis of plant STY protein kinases.

mRNA-binding protein HuR in breast carcinogenesis

Breast cancer is the most common cancer among women. Although its prognosis has improved nowadays, methods to predict the progression of the disease or to treat it are not comprehensive. This thesis work was initiated to elucidate in breast carcinogenesis the role of HuR, a ubiquitously expressed mRNA-binding protein that regulates gene expression posttranscriptionally. HuR is predominantly nuclear, but it shuttles between the nucleus and the cytoplasm, and this nucleocytoplasmic translocation is important for its function as a RNA-stabilizing and translational regulator. HuR has been associated with diverse cellular processes, for example carcinogenesis. The specific aims of my thesis work were to study the prognostic value of HuR in breast cancer and to clarify the mechanisms by which HuR contributes to breast carcinogenesis. My ultimate goal is, by better understanding the role of HuR in breast carcinogenesis, to aid in the discovery of novel targets for cancer therapies. HuR expression and localization was studied in paraffin-embedded preinvasive (atypical ductal hyperplasia, ADH, and ductal carcinoma in situ, DCIS) specimens as well in sporadic and familial breast cancer specimens. Our results show that cytoplasmic HuR expression was already elevated in ADH and remained elevated in DCIS as well as in cancer specimens. Clinicopathological analysis showed that cytoplasmic HuR expression associated with the more aggressive form of the disease in DCIS, and in cancer specimens it proved an independent marker for poor prognosis. Importantly, cytoplasmic HuR expression was significantly associated with poor outcome in the subgroups of small (2 cm) and axillary lymph node-negative breast cancers. HuR proved to be the first mRNA stability protein the expression of which is associated in breast cancer with poor outcome. To explore the mechanisms of HuR in breast carcinogenesis, lentiviral constructs were developed to inhibit and to overexpress the HuR expression in a breast epithelial cell line (184B5Me). Our results suggest that HuR mediates breast carcinogenesis by participating in processes important in cell transformation, in programmed cell death, and in cell invasion. Global gene expression analysis shows that HuR regulates genes participating in diverse cellular processes, and affects several pathways important in cancer development. In addition, we identified two novel target transcripts (connective tissue growth factor, CTGF, and Ras oncogene family member 31, RAB31) for HuR. In conclusion, because cytoplasmic HuR expression in breast cancer can predict the outcome of the disease it could serve in clinics as a prognostic marker. HuR accumulates in the cytoplasm even at its non-invasive stage (ADH and DCIS) of the carcinogenic process and supports functions essential in cell alteration. These data suggest that HuR contributes to carcinogenesis of the breast epithelium.

Crystal Structure of Staphylococcus aureus Metallopeptidase (Sapep) Reveals Large Domain Motions between the Manganese-bound and Apo-states

Proteases belonging to the M20 family are characterized by diverse substrate specificity and participate in several metabolic pathways. The Staphylococcus aureus metallopeptidase, Sapep, is a member of the aminoacylase-I/M20 protein family. This protein is a Mn2+-dependent dipeptidase. The crystal structure of this protein in the Mn2+-bound form and in the open, metal-free state suggests that large interdomain movements could potentially regulate the activity of this enzyme. We note that the extended inactive conformation is stabilized by a disulfide bond in the vicinity of the active site. Although these cysteines, Cys(155) and Cys(178), are not active site residues, the reduced form of this enzyme is substantially more active as a dipeptidase. These findings acquire further relevance given a recent observation that this enzyme is only active in methicillin-resistant S. aureus. The structural and biochemical features of this enzyme provide a template for the design of novel methicillin-resistant S. aureus-specific therapeutics.

Classification of Protein Kinases on the Basis of Both Kinase and Non-Kinase Regions

Background: Protein phosphorylation is a generic way to regulate signal transduction pathways in all kingdoms of life. In many organisms, it is achieved by the large family of Ser/Thr/Tyr protein kinases which are traditionally classified into groups and subfamilies on the basis of the amino acid sequence of their catalytic domains. Many protein kinases are multidomain in nature but the diversity of the accessory domains and their organization are usually not taken into account while classifying kinases into groups or subfamilies. Methodology: Here, we present an approach which considers amino acid sequences of complete gene products, in order to suggest refinements in sets of pre-classified sequences. The strategy is based on alignment-free similarity scores and iterative Area Under the Curve (AUC) computation. Similarity scores are computed by detecting common patterns between two sequences and scoring them using a substitution matrix, with a consistent normalization scheme. This allows us to handle full-length sequences, and implicitly takes into account domain diversity and domain shuffling. We quantitatively validate our approach on a subset of 212 human protein kinases. We then employ it on the complete repertoire of human protein kinases and suggest few qualitative refinements in the subfamily assignment stored in the KinG database, which is based on catalytic domains only. Based on our new measure, we delineate 37 cases of potential hybrid kinases: sequences for which classical classification based entirely on catalytic domains is inconsistent with the full-length similarity scores computed here, which implicitly consider multi-domain nature and regions outside the catalytic kinase domain. We also provide some examples of hybrid kinases of the protozoan parasite Entamoeba histolytica. Conclusions: The implicit consideration of multi-domain architectures is a valuable inclusion to complement other classification schemes. The proposed algorithm may also be employed to classify other families of enzymes with multidomain architecture.