Physiol. Genomics 28: 223-231, 2007.
First published October 10, 2006; doi:10.1152/physiolgenomics.00173.2006
1094-8341/07 $8.00
Received 5 August 2006;
accepted in final form 4 October 2006.
Physiological Genomics 28:223-231 (2007)
1094-8341/07 $8.00 © 2007 American Physiological Society
Genome-wide analysis of SPAK/OSR1 binding motifs
Eric Delpire and
Kenneth B. E. Gagnon
Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee
 |
ABSTRACT
|
|---|
Based on the alignment of 12 sequences of protein motifs that interact with the kinases SPAK (Ste20-related proline alanine-rich kinase) and OSR1 (oxidative stress response 1), we performed genome-wide searches of the sequence [S/G/V]RFx[V/I]xx[V/I/T/S]xx, where x represents any amino acid. The "Mus musculus" search resulted in the identification of 131 mouse proteins containing 137 SPAK/OSR1 putative binding motifs. Similar numbers were found for human, zebrafish, fruit fly, and worm. A little more than half of the mouse proteins containing SPAK/OSR1 binding domains (53%) were also identified in the human search, whereas 
17–18% of these common hits were identified in the zebrafish search. The mouse proteins could be divided into two broad categories: 2/3 had an identified function, whereas 1/3 were either predicted or of unknown function. The known proteins were grouped as transport proteins, other membrane proteins, kinases, phosphatases, cytoskeletal, ribosomal, nuclear, enzymes, and others. Analysis of the location of the SPAK/OSR1 binding motif within the protein sequence revealed distribution throughout the entire length, but with preference to the extreme amino- or carboxyl termini for a large number of proteins. Analysis of the amino acid composition of the motifs revealed a preponderance of serine residues at positions 5, 6, 7, and 8. In summary, our new search found and thus confirms the 12 proteins previously shown to interact with the kinases and identifies 119 potential new targets for SPAK and OSR1 in the mouse proteome.
mouse genome; protein-protein interaction; docking site; Ste20 kinases; NCBI search; Ste20-related proline alanine-rich kinase; oxidative stress response-1
|
INTRODUCTION
|
|---|
USING A LARGE-SCALE YEAST-2 hybrid screen and the amino terminus of KCC3 (K-Cl cotransporter-3), we uncovered a novel protein-protein interaction motif, recognized by two closely related Ste20 kinases: Ste20-related proline alanine-rich kinase (SPAK/PASK) and oxidative stress response-1 (OSR1) (5). Using deletion mutants and single residue mutagenesis, we determined that at minimum, nine residues in the target protein are required for kinase binding. The SPAK binding motif, published in 2002, has the sequence [R/K]Fx[V/I]xxxxx, where x represents any amino acid. Because the motif is based on yeast-2 hybrid data and this methodology cannot assess binding to residues located upstream of the positively charged arginine residue (that we will define as position 1), we cannot preclude the possibility that more than nine residues are involved in the interaction. Indeed, the smallest peptide (9 amino acids) that gives a positive interaction in a yeast-2 hybrid assay is fused to the binding domain of GAL4, and thus, amino acids linking this domain to the peptide might in fact participate to the binding. Methodologies different from those using fusion proteins (different than yeast-2 hybrid and glutathione-S-transferase pull-down) are needed to resolve the minimum length required for SPAK/OSR1 binding. Subsequent studies from our laboratory and others have demonstrated binding of SPAK to several additional proteins (3–6). Sequence alignment of the motifs found in these proteins is presented in Fig. 1.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 1. Alignment of the mouse KCC3 Ste20-related proline alanine-rich kinase/oxidative stress response-1 (SPAK/OSR1) binding motif with sequences of 11 other proteins experimentally shown to interact with the 2 kinases. Solid boxes delineate residues identified as essential for protein-protein interaction: R at position 1, F at position 2, and V/I at position 4. Dashed boxes identified residues exhibiting conservation at specific positions: V/S/G at position –1, and V/I/T/S at position 7.
|
|
Several conclusions can be made from this alignment: First, serine or glycine residues are found preferentially at position –1, with the only exception being the original KCC3 target, which contains a valine residue at that position. Second, despite our previous demonstration of a conserved protein-protein interaction between SPAK and NKCC2 when we substituted a lysine for the arginine residue (2), of the 12 proteins listed in Fig. 1, only arginine residues are found at position 1. Third, amino acid substitution at position 2 implies the necessity for a phenylalanine residue (2), a fact supported by the alignment represented in Fig. 1. Fourth, only valine or isoleucine residues are found at position 4. Finally, valine, isoleucine, threonine, and serine residues are found at position 7. Thus, based on this analysis, we propose the following refined motif: [S/G/V]RFx[V/I/]xx[V/I/T/S]xx. Since proteins are made of 20 amino acids, at first approximation, the probability of finding one specific residue at any position is 1 divided by 20 or 0.05. Thus, the probability of finding the [S/G/V]RFx[V/I/]xx[V/I/T/S]xx motif can be calculated to be 7.5 x 10–6. Because some amino acids are more abundant than others, the actual probability of finding the motif increases slightly to 8.46 x 10–6, indicating that if the SPAK binding motif was distributed within the genome by chance only, this motif should be found once every 120,000 (118,151) residues.
To determine the number of SPAK/OSR1 binding motifs existing within a genome and identify the proteins containing this motif, we analyzed the National Center for Biotechnology Information (NCBI) protein database. Because the search had to accommodate multiple residues at any specific position, we created a small program written in Visual Basic that allowed us to identify the binding motifs within entire genomes. We provide here a complete list of mouse proteins containing [S/G/V]RFx[V/I/]xx[V/I/T/S]xx motifs and provide information regarding conservation between different vertebrate genomes.
|
METHODS
|
|---|
The NCBI protein database(http://www.ncbi.nlm.nih.gov/) was searched for Mus musculus, and hits were saved in FASTA format within a single file. After opening the file and removing all "Hard Returns" using the find and replace function of WordPerfect, we saved the file in TEXT format. Next, using a small routine written in Visual Basic (Microsoft), we searched the entire text file for specific sequences allowing multiple residues per position. Basically, the first 60 characters are placed into 60 consecutive string variables, and the routine analyzes the file one variable at a time. Based on the identity of the first variable, a specific action is taken (Fig. 2), and the routine moves one character over, resetting all 60 variables until the end of the file is reached. All NCBI FASTA entries begin with a line of text describing the sequence (i.e., >gi 51592076 ref NP_032089.2 solutecarrierfamily37). Our program identifies new protein entries when the first variable = ">", the second = "g", and the third = "i". The routine then increments its protein counter by one and seeks the fourth "|" character, which signals the beginning of the name or description of the protein. The 60 characters following the fourth "|" character are saved temporarily in a string variable. Next, the routine seeks 30 consecutive capitalized letters, which signals the beginning of the sequence (this method was chosen since all NCBI entries do not start with a methionine), and starts incrementing its residue counter. When the routine identifies a serine, valine, or glycine at the first position (first variable), followed by arginine and phenylalanine residues at positions 2 and 3, respectively, followed by a valine or isoleucine at position 5, and finally followed by valine, isoleucine, serine, or threonine residues at position 8, the routine captures the eight characters plus an additional seven characters following the motif into a temporary variable. This variable is then compared with 15 letter motifs previously saved, and if the string of 15 characters is unique, the motif belongs to a new protein and is saved along with the variable containing the name of the protein, and the motif counter is incremented by one. In contrast, if the string of 15 character is already saved, this indicates a redundant protein and the "duplicate protein" counter is incremented by one. The likelihood that two different proteins contain the exact same 15 characters is infinitesimal, as the probability to find a specific string of 15 residues can be estimated at 3 x 10–20. The routine then continues its search without saving the motif. At the end of the search, all variables (protein counter, motif counter, protein names, and motifs) are copied into a new single text file.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 2. Logic tree describing the general process used by the Visual Basic routine to identify SPAK/OSR1 binding motifs. See METHODS for details.
|
|
For each of the proteins returned by the Visual Basic routine, we queried the NCBI protein database to locate and confirm the motif within the amino acid sequence of the protein. Additionally, the curation process provided the opportunity to eliminate any remaining redundant or nonmurine proteins and allowed grouping of the proteins into categories based on their general description and features.
|
RESULTS AND DISCUSSION
|
|---|
The NCBI protein database(http://www.ncbi.nlm.nih.gov/) was first searched for Mus musculus. The search was performed on June 14, 2006, and returned some 189,153 hits. More directed searches such as Mus musculus AND "descriptor" returned: 21,389 hits for "receptors," 8,228 hits for "kinases," 3,018 hits for "enzymes," 1,389 hits for "cytoskeleton," 3,331 hits for "transport," 2,412 hits for "phosphatases," 2,601 hits for "channels." Further defining the type of ion channel reduced the number of hits to 551 for "K+ channels," 307 for "Ca2+ channels," 272 hits for "Na+ channels," and 184 hits for "Cl– channels" (see Table 1). An obvious first observation that can be made from these searches is that NCBI returns many more hits than the number of proteins existing in any one specific genome. Indeed, the NCBI search of Mus musculus returned close to 200,000 hits, whereas estimation of the number of proteins in the mouse genome is closer to 24,000–25,000 (7). A second observation is that, due to the extensive cross-referencing of NCBI entries, proteins are often identified using unrelated or indirect searches. As an example, only a fraction of the 8,228 entries found with Mus musculus AND "kinases" are actual kinases. The Na-K-2Cl cotransporter (NKCC1) is identified in the Mus musculus AND "kinases" search due to the multiple NCBI links existing between SPAK/PASK kinases and the cotransporter.
Using a routine written in Visual Basic, we searched for occurrence of the [S/G/V]RFx[V/I]xx[V/I/T/S]xx within the 189,153 downloaded mouse proteins. The routine identified 170 proteins containing 175 motifs. Analysis of the proteins showed that the routine was able to eliminate all redundant entries and alternatively spliced products, thus identifying one product per gene. However, a small number of proteins identified in the "Mus musculus" search were not mouse proteins but belonged to other species. After curation and elimination of these proteins the final list comprised 131 proteins containing 137 motifs. There were five instances where a single open reading frame encoded a protein containing more than one SPAK/OSR1 binding motifs. WNK1 contains three independent SPAK/OSR1 binding motifs, and four other proteins (NKCC1, AATYK, titin isoform N2B, and mKIAA1297) contain two SPAK/OSR1 motifs. The complete list is presented in Table 2. Two-thirds of these proteins are involved in transport, enzymatic, cytoskeletal, nuclear, ribosomal, and other functions, whereas the remaining one-third of the proteins listed in Table 2 either are only predicted or have undetermined function.
Comparison among genomes revealed a high number of proteins with conserved SPAK/OSR1 binding motifs. Indeed, 70 out of the 131 mouse proteins were also found in the human genome, indicating a 53% conservation. However, a significantly lower number of mouse and human proteins containing SPAK/OSR1 binding motifs (23 out of 131 mouse proteins) were present in the genetically more distant zebrafish genome, corresponding to 17.5% of mouse proteins. These data are summarized in Table 2 and Fig. 3.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 3. Conservation of SPAK/OSR1 binding motifs. The value in parentheses indicates the number of protein hits in 3 vertebrate genomes: mouse (Mus musculus), human (Homo sapiens), and zebrafish (Danio rerio). We identified a total of 131 mouse proteins from the 170 hits. There were 23 out of 131 proteins conserved in mouse, human, and zebrafish genomes. Sixty-one proteins were specific to the mouse genome, and 70 (47 + 23) proteins were conserved between the mouse and human genomes. The list of these proteins is reported in Table 2.
|
|
Next, we examined the location of the SPAK/OSR1 binding motif within the entire set of 131 proteins. As seen in Fig. 4, a larger number of motifs are located at the extreme NH2 terminus of the protein. The remaining motifs are distributed along the length of the target protein. However, proteins such as transporters and channels, which have a transmembrane core and cytosolic amino- and carboxyl termini, have the SPAK/OSR1 binding motifs preferentially located on the cytosolic tails (see Table 2). In other proteins such as kinases, which have well-defined catalytic and regulatory domains, the [S/G/V]RFx[V/I]xx[V/I/T/S]xx motifs are found exclusively in the regulatory domain. Indeed, the average location for sites found in kinases displaying SPAK/OSR1 binding motifs is 80 ± 3.2% (n = 11). With the exception of dynein, all SPAK/OSR1 binding motifs found in cytoskeletal proteins are also found at the extreme carboxyl-terminus (Table 2). Finally, note the opposite location of the SPAK/OSR1 binding motifs in the two regulatory subunits of protein phosphatase 2 (Table 2). These motifs are again preferentially located at the very beginning or very end of the protein.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4. Distribution of the SPAK/OSR1 motifs within the length of proteins. For each motif identified, its location (position –1) was divided by the total length of the protein. There were 137 motifs identified in the mouse search. Note that the motif is distributed throughout the length of the protein with a larger number of proteins with SPAK/OSR1 binding motif located within the 1st 10% of the protein length. The motif location for each individual protein is reported in Table 2.
|
|
The identification of 137 [S/G/V]RFx[V/I]xx[V/I/T/S]xx motifs within 131 mouse proteins allow us to further examine the distribution of residues in the motif at positions 3 and 5–9. As shown in Fig. 5, residues at position 3 can be very diverse, from a mostly hydrophobic leucine to a very hydrophilic arginine. One interesting feature of this analysis of the expanded motif, is the preferential use of serine residues at positions 5, 6, 7, and 8.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 5. Relative frequency of occurrence of amino acids at each position of the SPAK/OSR1 binding motif. Analysis was based on the 137 binding motifs listed in Table 2. Size of the letter is proportional to the occurrence of the amino acid. The arginine residue was arbitrarily placed at position 1.
|
|
The list of proteins that contain the SPAK/OSR1 binding motif is rather extensive despite the somewhat restrictive definition of the motif based on the 12 sequences reported in Fig. 1. However, if we were to allow additional residues at positions –1, 1, or 7, the total number of proteins would increase significantly. For example, we expanded our search to include the possibility of a lysine residue at position 1. We identified an additional 100 proteins with [S/G/V]KFx[V/I] xx[V/I/T/S]xx motifs (Table 3). To date, however, no SPAK/OSR1 interacting proteins have been identified as containing a KFxV motif. As the motifs were identified through gene searching and not from peptide libraries, some of the identified motifs might not constitute actual SPAK/OSR1 binding sites. Therefore, whether proteins identified in this search do indeed interact with the kinases requires experimental testing. Anchoring of kinases to their substrates constitutes one important mechanism of substrate specificity (1). Thus, for any protein-protein interaction to reach biological significance, overlap of their temporal and spatial distribution is required. While SPAK/OSR1 interaction with cation-chloride cotransporters, chloride channels, and TNF receptor has been evidenced experimentally (3–6), this study provides a list of new proteins that potentially interact with SPAK and OSR1 and suggests a significantly greater physiological role for each of these kinases.
|
GRANTS
|
|---|
This work was supported by National Institutes of Health Research Grant NS-36758.
|
FOOTNOTES
|
|---|
Address for reprint requests and other correspondence: E. Delpire, Dept. of Anesthesiology, Vanderbilt Univ. Medical Ctr., T-4202 Medical Center No., 1161 21st Ave. S., Nashville, TN 37232 (e-mail: eric.delpire{at}vanderbilt.edu).
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
|
REFERENCES
|
|---|
- Bardwell AJ, Flatauer LJ, Matsukuma K, Thorner J, Bardwell L. A conserved docking site in MEKs mediates high-affinity binding to MAP kinases and cooperates with a scaffold protein to enhance signal transmission. J Biol Chem 276: 10374–10386, 2001.[Abstract/Free Full Text]
- Delpire E, Piechotta K. Ste20 kinases and cation-chloride cotransporters. In: Cell Volume and Signal Transduction, edited by Lauf PK and Adragna NC (Advances in Experimental Medicine and Biology). New York: Springer Science and Business Media 559: 43–53, 2004.
- Denton J, Nehrke K, Yin X, Morrison R, Strange K. GCK-3, a newly identified Ste20 kinase, binds to and regulates the activity of a cell cycle-dependent ClC anion channel. J Gen Physiol 125: 113–125, 2005.[Abstract/Free Full Text]
- Piechotta K, Garbarini NJ, England R, Delpire E. Characterization of the interaction of the stress kinase SPAK with the Na+-K+-2Cl– cotransporter in the nervous system: Evidence for a scaffolding role of the kinase. J Biol Chem 278: 52848–52856, 2003.[Abstract/Free Full Text]
- Piechotta K, Lu J, Delpire E. Cation-chloride cotransporters interact with the stress-related kinases SPAK and OSR1. J Biol Chem 277: 50812–50819, 2002.[Abstract/Free Full Text]
- Polek TC, Talpaz M, Spivak-Kroizman T. The TNF receptor, RELT, binds SPAK and uses it to mediate p38 and JNK activation. Biochem Biophys Res Commun 343: 125–134, 2006.[CrossRef][ISI][Medline]
- Waterston RH, Mouse Genome Sequencing Consortium. Initial sequencing and comparative analysis of the mouse genome. Nature 420: 520–562, 2002.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
C. Richardson and D. R. Alessi
The regulation of salt transport and blood pressure by the WNK-SPAK/OSR1 signalling pathway
J. Cell Sci.,
October 15, 2008;
121(20):
3293 - 3304.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. Ponce-Coria, P. San-Cristobal, K. T. Kahle, N. Vazquez, D. Pacheco-Alvarez, P. de los Heros, P. Juarez, E. Munoz, G. Michel, N. A. Bobadilla, et al.
Regulation of NKCC2 by a chloride-sensing mechanism involving the WNK3 and SPAK kinases
PNAS,
June 17, 2008;
105(24):
8458 - 8463.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. Richardson, F. H. Rafiqi, H. K. R. Karlsson, N. Moleleki, A. Vandewalle, D. G. Campbell, N. A. Morrice, and D. R. Alessi
Activation of the thiazide-sensitive Na+-Cl- cotransporter by the WNK-regulated kinases SPAK and OSR1
J. Cell Sci.,
March 1, 2008;
121(5):
675 - 684.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. Uvarov, A. Ludwig, M. Markkanen, P. Pruunsild, K. Kaila, E. Delpire, T. Timmusk, C. Rivera, and M. S. Airaksinen
A Novel N-terminal Isoform of the Neuron-specific K-Cl Cotransporter KCC2
J. Biol. Chem.,
October 19, 2007;
282(42):
30570 - 30576.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
K. B. E. Gagnon, R. England, L. Diehl, and E. Delpire
Apoptosis-associated tyrosine kinase scaffolding of protein phosphatase 1 and SPAK reveals a novel pathway for Na-K-2C1 cotransporter regulation
Am J Physiol Cell Physiol,
May 1, 2007;
292(5):
C1809 - C1815.
[Abstract]
[Full Text]
[PDF]
|
|
|
Copyright © 2007 by the American Physiological Society.
TOP
ABSTRACT
INTRODUCTION
