A versatile mass spectrometry-based method to both identify kinase client-relationships and characterize signaling network topology

Abstract

While more than a thousand protein kinases (PK) have been identified in the Arabidopsis thaliana genome, relatively little progress has been made toward identifying their individual client proteins. Herein we describe the use of a mass spectrometry-based in vitro phosphorylation strategy, termed Kinase Client assay (KiC assay), to study a targeted-aspect of signaling. A synthetic peptide library comprising 377 in vivo phosphorylation sequences from developing seed was screened using 71 recombinant A. thaliana PK. Among the initial results, we identified 23 proteins as putative clients of 17 PK. In one instance protein phosphatase inhibitor-2 (AtPPI-2) was phosphorylated at multiple-sites by three distinct PK, casein kinase1-like 10, AME3, and a Ser PK-like protein. To confirm this result, full-length recombinant AtPPI-2 was reconstituted with each of these PK. The results confirmed multiple distinct phosphorylation sites within this protein. Biochemical analyses indicate that AtPPI-2 inhibits type 1 protein phosphatase (TOPP) activity, and that the phosphorylated forms of AtPPI-2 are more potent inhibitors. Structural modeling revealed that phosphorylation of AtPPI-2 induces conformational changes that modulate TOPP binding.

Figure 1

Experimental strategy used for identification of kinase-client relationships.

Figure 2

Comparative phosphosites mapping of AtPPI-2 (At5g52200). Experimentally-determined, in vivo phosphorylation sites for AtPPI-2 from P³DB and PhosPhAt database, compared with sites obtained in vitro (KiC assay) using recombinant AtPPI-2 and PKs (CK1 like-10, AME3 and Ser PK like protein) (A). Letters in gray indicates low-confidence phosphosites. B, represent a MS/MS spectrum of a novel phosphosite in AtPPI-2. The phosphopeptide NVLNDAAASsSR represent an example of an additional novel phosphosite (Ser⁷⁷) of the AtPPI-2. Small letter in the sequence indicating the phosphosite. Arrow and asterisks indicates the detection of the neutral loss of the phosphate on fragments or precursor ions and the phosphorylated Ser residues, respectively.

Figure 3

Comparison of the in vitro phosphorylation-site specificities of recombinant CK1-like 10, AME3, and Ser PK like protein. In separate experiments, recombinant AtPPI-2 was phosphorylated with an individual kinase, digested with trypsin, and the tryptic peptides analyzed by MS. Data presented are means of three biological replicates. Percentage shown above the bars indicates the pRS site probability for each amino acid. Based on the pRS score and pRS site probability (Table S5) solid and empty bars represent the phosphosites with high and low confidence, respectively.

Figure 4

Phosphorylation of AtPPI-2 makes it a better inhibitor of λ TOPP activity. Recombinant AtPPI-2 was incubated with λ TOPP either before or after phosphorylation by the indicated kinases. TOPP activity was measured at an absorbance A₄₀₅ wherein p-nitrophenyl phosphate (pNPP) was used as substrate. Inhibotory activity is expressed as a percentage of the control values. Values are means ± SEM for control (n = 9) or plus AtPPI-2 (n = 3) reactions. Different letters above the bars indicate a statistically-significant difference (p<0.05) according to the Duncan multiple-range test.

Figure 5

The topological relationships among A. thaliana seed-expressed PK and their cognate client-proteins. The data were obtained by in vitro screening of a synthetic peptide library derived from in vivo phospho-proteomic analyses of developing A. thaliana and B. napus seeds (P³DB) with 71 recombinant A. thaliana PK. Based upon the peptide sequences, 23 proteins were identified as presumptive clients for 17 PK. The cartograph was assembled using the organic layout from Cytoscape version 2.3.2 . The PKs that belong to the same family are presented as hubs of the same color, while the client nodes are uncolored. Each edge represents a single PK-client pair. The spatial distribution of hubs, nodes, and edges is arbitrary.

Figure 6

Predicted conformational changes in the AtPPI-2 structure induced by phosphorylation of Ser¹⁴⁰. Panel (A) includes the model of the rat TOPP catalytic domain 1 subunit (cyan, PDB accession 2O8AA0) docked with mouse PPI-2 (purple, PDB accession 2O8AI0). Panels B, C and D contain the model of the rat TOPP catalytic domain docked with non-phosphorylated (B) or phosphorylated (C and D) AtPPI-2, respectively. Protein phosphorylation was mimicked by Ser to Glu (C) and Asp (D) changes. Phosphorylation of Ser¹⁴⁰ cause conformational changes in AtPPI-2. Panels E, F, G and H include close-up views of the interaction between the conserved RKxHY motif of mouse PPI-2 (E) and/or AtPPI-2 (F–G) with residues Arg⁹⁶, His¹²⁵, Arg²²¹, and Tyr²⁷² of TOPP. The AtPPI-2 binding sites in TOPP are colored magenta. Structures were produced using PyMOL (www.pymol.org).

Figure 7

Analysis of the topology of signaling network interactions. This model shows AtPPI-2 located at the interface of multiple signaling networks. Both inactive and activated (phosphorylated) forms of AtPPI-2 bind to TOPP, inhibiting protein-phosphatase activity. The phosphorylated form of AtTOPP-2 is, however, a more effective protein-phosphatase inhibitor. In this model, activation of AtPPI-2 can be mediated through phosphorylation by three separate kinases (CK 1-like 10, a LAMMER-type protein kinase (AME3), or a Ser/Thr PK-family protein), any pair of these kinases, or all three kinases. In all instances, phosphorylation of PPI-2 would result from transduction of a signal from the affiliated network. A, B, and C represent phosphorylated residues. In the un-phosphorylated form they are represented by –OH groups, while in the activated form of AtPPI-2 they are designated by –P. A (CK 1-like 10) includes Ser⁴⁵, Ser⁷⁷, Ser¹⁴⁰, and Thr¹⁷⁸; B (AME3), Ser⁷⁷, Ser¹⁴⁰, and Ser¹⁵⁵; and C (S/T PK-like), Ser⁴⁵, Ser¹⁴⁰, and Thr¹⁷⁸. This model depicts the complexity of interactions (multiple networks, kinases, and phosphorylation-sites) that might be controlled from a single-interface among pathways.