On the substrate recognition and negative regulation of SPAK, a kinase modulating Na+-K+-2Cl- cotransport activity

Abstract

Threonines targeted by Ste20-related proline-alanine-rich kinase (SPAK) for phosphorylation have been identified in Na+-K+-2Cl(-) cotransporter type 1 (NKCC1), NKCC2, and Na+-Cl(-) cotransporter (NCC). However, what constitutes the substrate recognition of the kinase is still unknown. Using site-directed mutagenesis and functional measurement of NKCC1 activity in Xenopus laevis oocytes, we determined that SPAK recognizes two threonine residues separated by four amino acids. Addition or removal of a single residue abrogated SPAK activation of NKCC1. Although both threonines are followed by hydrophobic residues, in vivo experiments have determined that SPAK activation of the cotransporter only requires a hydrophobic residue after the first threonine. Interestingly, downstream of the second threonine residue, we have identified a conserved aspartic acid residue which is critical for NKCC1 function. Mouse SPAK activity requires phosphorylation of two specific residues by WNK [with no lysine (K)] kinases: a threonine (T243) in the catalytic domain and a serine (S383) in the regulatory domain. We found that mutating the threonine residue into a glutamic acid (T243E) combined with mutation of the serine into an aspartic acid (S383D) rendered SPAK constitutively active. Surprisingly, alanine substitution of S383 or mutation of residues surrounding this residue also resulted in a constitutively active kinase. Interestingly, deletion of amino acids 356-398 identified another serine residue in the catalytic domain (S321) as another putative target of WNK phosphorylation. We found that WNK4 is capable of stimulating the deletion mutant when S321 is present, but not when S321 is mutated into an alanine.

Fig. 1.

Alignment of mouse and human Na⁺-dependent cation chloride cotransporter sequences around the phosphothreonines. Identical residues are in red font over yellow background, while conserved residues are in black or blue fonts over green and blue backgrounds. The sites of phosphorylation are indicated with a star. Note that the residues that directly follow the threonine residues are hydrophobic (ϕ) in nature. Numbers at top indicate amino acid position of mouse Na-K-2Cl cotransporter type 1 (NKCC1). NCC, Na⁺-Cl⁻ cotransporter.

Fig. 2.

Distance between T206 and T211 affects NKCC1 activity in Xenopus laevis oocytes. Cotransporter activity was measured through K⁺(⁸⁶Rb⁺) uptake under isosmotic (white bars), isosmotic with Ste20-related proline-alanine-rich kinase (SPAK) and WNK4 [with no lysine (K)] coexpression (gray bars), or hyperosmotic (black bars) conditions. Residue numbers correspond to the mouse NKCC1 sequence. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was reproduced at least once with independent oocytes from different frogs. All underlined mutants were significantly (P < 0.001) different from control (wild-type NKCC1) K⁺ uptake under their respective conditions (see materials and methods). Inset: biotinylated NKCC1 signal obtained with different nonfunctional cotransporters, compared with wild-type NKCC1.

Fig. 3.

Additional mutations around T206 and T211 on NKCC1 activity in Xenopus laevis oocytes. NKCC1 activity was measured through K⁺(⁸⁶Rb⁺) uptake under isosmotic (white bars), isosmotic with SPAK and WNK4 coexpression (gray bars), or hyperosmotic (black bars) conditions. Residue numbers correspond to the mouse NKCC1 sequence. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was reproduced at least once with independent oocytes from different frogs. *Significant (P < 0.001) difference from control (wild-type NKCC1) K⁺ uptake under their respective conditions. The underlined mutant was also significantly (P < 0.001) different from control (wild-type NKCC1) K⁺ uptake under their respective conditions (see materials and methods). Inset: biotinylated NKCC1 signal obtained with F207A cotransporter compared with wild-type NKCC1.

Fig. 4.

Role of two conserved aspartic acids on NKCC1 activity in Xenopus laevis oocytes. NKCC1 function was measured through K⁺(⁸⁶Rb⁺) uptake under isosmotic (white bars), isosmotic with SPAK and WNK4 coexpression (gray bars), or hyperosmotic (black bars) conditions. Residue numbers correspond to the mouse NKCC1 sequence. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was reproduced at least once with independent oocytes from different frogs. Underlined mutants were significantly (P < 0.001) different from control (wild-type NKCC1) K⁺ uptake under their respective conditions (see materials and methods). Inset: biotinylated NKCC1 signal obtained with different nonfunctional cotransporters, compared with wild-type NKCC1.

Fig. 5.

Targeted mutations in SPAK and oxidative stress-responsive kinase 1 (OSR1) render the kinases constitutively active. Kinase activity was assessed in Xenopus laevis oocytes through its ability to activate wild-type NKCC1 in the absence (white bars) or presence (black bars) of WNK4. NKCC1 activity was measured through K⁺(⁸⁶Rb⁺) uptake under isosmotic conditions. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was tested at least twice with independent oocytes isolated from different frogs. *Significant (P < 0.001) difference from control conditions (NKCC1 + wild-type SPAK or NKCC1 + wild-type SPAK + WNK4).

Fig. 6.

Targeted mutations in the S383-containing segment activate SPAK 243E. To test for constitutive activity in the absence of WNK4, NKCC1 activity was measured through K⁺(⁸⁶Rb⁺) uptake in Xenopus laevis oocytes coexpressing SPAK with the T243E mutation only (wild-type) or containing additional mutations in the S383 domain. All fluxes were performed under isosmotic conditions. The residues surrounding S383 and targeted for mutation are presented. S383 is denoted by an asterisk. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was tested at least twice with independent oocytes isolated from different frogs. All K⁺ uptakes with mutant kinases were significantly (P < 0.001) different from K⁺ uptake with wild-type kinase.

Fig. 7.

Role of S321 in SPAK activation. Top: schematic representation of SPAK showing the position of the putative residues phosphorylated by WNK4. Horizontal line represents deleted segment (amino acid residues 356–398). Boxed residues (DEMD) represent a caspase cleavage site. Bottom: effect of deletion of a 43 amino acid fragment of the regulatory domain on SPAK activity. Kinase activity was assessed in Xenopus laevis oocytes through its ability to activate NKCC1 in the presence or absence of WNK4. NKCC1 activity was measured through K⁺(⁸⁶Rb⁺) uptake under isosmotic conditions. Inset: sequence surrounding the S/T residues which are putative targets of WNK4 phosphorylation. Basic residues at positions P-2 and P-5 are colored red and underlined. Bars represent means ± SE of the flux (expressed in nanomoles per oocyte per hour; n = 20–25 oocytes). Each condition was tested at least twice with independent oocytes isolated from different frogs. *Significant (P < 0.001) difference between wild-type NKCC1 coexpressed with wild-type and mutant forms of SPAK.