Chemical inhibition of prometastatic lysyl-tRNA synthetase-laminin receptor interaction

Abstract

Lysyl-tRNA synthetase (KRS), a protein synthesis enzyme in the cytosol, relocates to the plasma membrane after a laminin signal and stabilizes a 67-kDa laminin receptor (67LR) that is implicated in cancer metastasis; however, its potential as an antimetastatic therapeutic target has not been explored. We found that the small compound BC-K-YH16899, which binds KRS, impinged on the interaction of KRS with 67LR and suppressed metastasis in three different mouse models. The compound inhibited the KRS-67LR interaction in two ways. First, it directly blocked the association between KRS and 67LR. Second, it suppressed the dynamic movement of the N-terminal extension of KRS and reduced membrane localization of KRS. However, it did not affect the catalytic activity of KRS. Our results suggest that specific modulation of a cancer-related KRS-67LR interaction may offer a way to control metastasis while avoiding the toxicities associated with inhibition of the normal functions of KRS.

Figure 1. Eeffects of KRS on cancer metastasis

(a–d) Established stable mouse breast carcinoma 4T1 cells with high expression of KRS (KRS-1 and KRS-2) or low expression of KRS (sh-KRS-1 and sh-KRS-2) were injected subcutaneously on the back of BALB/cAnCr mice (n = 5). After we sacrificed the mice, we assessed cancer metastasis by pulmonary nodule formation (scale bar = 0.5 cm) (a, c). The numbers of pulmonary nodules (>1 mm in size) of stable KRS-1/2 cells (b) and stable sh-KRS-1/2 cells (d) were counted and presented as mean ± s.d. (b: ** P < 0.01; d: *** P < 0.001).

Figure 2. Identification of YH16899 inhibiting KRS–LR interaction and cell invasion

(a) A schematic diagram of the screening strategy for an Y2H assay. KRS and 37LRP were fused to a DNA-binding domain (BD, LexA) and an activating domain (AD, B42) of a transcription factor, respectively. Yeast cells co-expressing these fusion proteins survive because BD and AD are brought in proximity via the KRS–LR interaction initiating transcription of a reporter gene (left). If chemicals interfere with the KRS–LR interaction, transcription of the reporter gene cannot occur and yeast cells cannot grow (right). (b) Chemical structure of the initial compound BC-K01 and its derivative YH16899. (c) H226 cells were incubated with YH16899, and KRS was immunoprecipitated for the detection of coprecipitated 67LR level using immunoblotting; WCL, whole cell lysate (See Supplementary Fig. 24 for raw data). (d) KRS-transfected H226 cells were treated with YH16899 (50 μM). Cell morphology was monitored using phase contrast (P.C.) microscopy (left). Actin and FAK activity were determined using immunofluorescence staining with rhodamine phalloidin (red, middle) and anti–phospho-Y397 (pY³⁹⁷)-FAK antibody (green, right), respectively (scale bar = 10 μm). (e) H226 cells were subjected to a Matrigel invasion assay in the presence of YH16899, and the number of cells penetrating the membrane were counted and presented as mean ± s.d. (f) H226 cells and YH16899 (50 μM) mixture were injected into the upper layer of CAM and incubated to promote invasion. The amount of human Alu gene was presented as mean ± s.d. (n = 3; *** P < 0.001).

Figure 3. Mapping and validation of an YH16899 docking site

(a) A schematic of the functional domain arrangement in human KRS. KRS contains an N-terminal extension domain (N-ext), an anticodon-binding domain (ABD), and a catalytic domain (CD). The T52 residue (undergoes phosphorylation) is indicated as a red dot. N-ext and ABD of KRS specifically interact with LR and YH16899. Arrow width indicates intensity of the compound–KRS interaction. (b) Spatial arrangement of the chemically perturbed residues (red) on KRS with tRNA^Lys. (c) Docking positions of the R- (left) and S- (right) YH16899 in the hydrophobic pocket of KRS were identified in NMR experiments. Ten low-energy positions were represented, and the CSP-defined residues are shown in red. (d) A summary of the critical residues around the YH16899 binding site. Specific residues were replaced with alanine, and the effects of the mutations on the binding to YH16899 and LR (in vitro pull-down and co-IP), 67LR stability, and cell invasion were examined (see Supplementary Fig. 18). The results of these experiments were calculated based on the relative band intensities (pull-down, IP, and 67LR level) and the relative numbers of migratory cells (invasion assay), and presented as a heat map with the scale from 0 to 10. For the SPR, a –log-transformed value of each K_D was obtained and presented on a scale from 0 (K_D, 1 × 10⁻³) to 10 (K_D, 1 × 10⁻¹⁰). The location of L142 and F144 in ABD, the most critical residues, is denoted in Fig. 3c.

Figure 4. Effects of YH16899 binding on the KRS structure

(a) SAXS and HDX-MS analysis of the effect of YH16899 on conformational changes of dimeric KRS T52D. KRS T52D has a more extended N-terminus and open ABD-CD interface compared to WT KRS, partially exposing the AIMP2-binding site. Incubation of KRS T52D with YH16899 induced a conformational change folding N-ext back toward the body of KRS. Blue and red indicate a less and more exposed surface, respectively. KRS WT was used as a control. (b) HDX-MS analysis of the effect of YH16899 on conformational changes of KRS T52D. Fragments of both N-ext and ABD showed decreased deuterium uptake in T52D + YH16899 compared to T52D alone. Cyan, 20%–40% decrease; Blue, >40% decrease. The mutagenesis-mapped binding site of inhibitor is colored in red. (c) CSP of [¹⁵N]KRS_1–72 by means of phospholipid nanodisc binding. Superposition of the 2D ¹H-¹⁵N TROSY spectra of free 0.2 mM KRS_1–72 with 0 mM (black), 0.1 mM (blue), and 0.2 mM (red) of the nanodisc. Strongly perturbed residues are denoted as blue circles.

Figure 5. A schematic model for the mode of action of YH16899

Phosphorylation of KRS at T52 induces a conformational change that opens up the AIMP2-binding interface, releasing KRS from MSC. YH16899 binds to KRS ABD (blue) and N-ext (grey) thereby directly inhibiting the interaction between KRS and 67LR. In addition, YH16899’s binding to KRS reduces the flexibility of N-ext and reduces the membrane localization of KRS. 67LR in the membrane without binding to KRS undergoes ubiquitin-mediated degradation. Although 67LR is thought to be a dimer of 37LRP, only a monomer is shown for the sake of simplicity.