Oligomeric States and Hydrodynamic Properties of Lysyl Oxidase-Like 2

Abstract

Lysyl oxidase-like 2 (LOXL2) has emerged as a promising therapeutic target against metastatic/invasive tumors and organ and tissue fibrosis. LOXL2 catalyzes the oxidative deamination of lysine and hydroxylysine residues in extracellular matrix (ECM) proteins to promote crosslinking of these proteins, and thereby plays a major role in ECM remodeling. LOXL2 secretes as 100-kDa full-length protein (fl-LOXL2) and then undergoes proteolytic cleavage of the first two scavenger receptor cysteine-rich (SRCR) domains to yield 60-kDa protein (Δ1-2SRCR-LOXL2). This processing does not affect the amine oxidase activity of LOXL2 in vitro. However, the physiological importance of this cleavage still remains elusive. In this study, we focused on characterization of biophysical properties of fl- and Δ1-2SRCR-LOXL2s (e.g., oligomeric states, molecular weights, and hydrodynamic radii in solution) to gain insight into the structural role of the first two SRCR domains. Our study reveals that fl-LOXL2 exists predominantly as monomer but also dimer to the lesser extent when its concentration is <~1 mM. The hydrodynamic radius (R_h) determined by multi-angle light scattering coupled with size exclusion chromatography (SEC-MALS) indicates that fl-LOXL2 is a moderately asymmetric protein. In contrast, Δ1-2SRCR-LOXL2 exists solely as monomer and its R_h is in good agreement with the predicted value. The R_h values calculated from a 3D modeled structure of fl-LOXL2 and the crystal structure of the precursor Δ1-2SRCR-LOXL2 are within a reasonable margin of error of the values determined by SEC-MALS for fl- and Δ1-2SRCR-LOXL2s in mature forms in this study. Based on superimposition of the 3D model and the crystal structure of Δ1-2SRCR-LOXL2 (PDB:5ZE3), we propose a configuration of fl-LOXL2 that explains the difference observed in R_h between fl- and Δ1-2SRCR-LOXL2s in solution.

Keywords: analytical ultracentrifugation; extracellular matrix; hydrodynamic radius; isoelectric points; lysyl oxidase-like 2; scavenger receptor cysteine-rich.

Figure 1

A schematic diagram of rLOXL2s in this study. fl-LOXL2: full-length LOXL2; Δ1-2SRCR-LOXL2: LOXL2 which lacks the first two SRCR domains; Δ1-3SRCR-LOXL2: LOXL2 which lacks the first three SRCR domains; Δ1-4SRCR-LOXL2: LOXL2 which lacks all four SRCR domains.

Figure 2

X-ray structure of Δ1-2SRCR-LOXL2 in the precursor form (PDB: 5ZE3). (A) Monomer (B) The active site structure. The precursor residues (Lys653, Tyr689) are 16.6 Å apart. Zn²⁺ (gray sphere) occupies the predicted Cu²⁺ binding site (His626-X-His628-X-His630). Ca²⁺ is shown as a green sphere. (C) In each asymmetric unit (ASU), two molecules are observed. The intermolecular association between monomers is mediated mainly through the 4th SRCR domains (in forest and light green). (D) A top view of the ASU shown in (C). (E) The 3rd SRCR domain (in cyan) interacts with the catalytic domain (in yellow) mostly through hydrogen bonding interactions (in red) but also through van der Waals interactions (in green). (F) The intermolecular association observed in between the 4th SRCR domains (shown in C,D) are all through van der Waals interactions (in green). Numbers in (A,E,F) are all in Å.

Figure 3

Polyacrylamide gel electrophoresis (PAGE) of LOXL2s. (A) SDS-PAGE under denaturing conditions in the presence and absence of reductant (β-mercaptoethanol). (B) Native-PAGE. AAQ: R315A/R316A/K317Q mutant form of LOXL2; WT: wild type; K317R-LOXL2 was isolated as a mixture of fl-LOXL2 and Δ1-2SRCR-LOXL2; Δ1-2SRCR-LOXL2 was isolated from the mixture by FPLC; r-P4H: recombinant Bacillus anthracis prolyl-4-hydroxylase as a molecular standard (homodimer, 49.2 kDa) [22].

Figure 4

Molecular weight (MW) determination of LOXL2 by SEC-MALS (A) fl-LOXL2 (B) Δ1-2SRCR-LOXL2; for clarity only every 10th measurement of MW is plotted.

Figure 5

The oligomeric state of fl-LOXL2. (A) A representative run of SV-AUC with 0.88 mg/mL of fl-LOXL2. (B) A digitized graph showing the percentage of monomer, dimer, tetramer and pentamer of fl-LOXL2.

Figure 6

Hydrodynamic radii (R_h) of fl-LOXL2: (A) Δ1-2SRCR-LOXL2 (B) calculated from SEC-MALS analysis.

Figure 7

Structure prediction of fl-LOXL2. (A) Overlay of 3D structure of fl-LOXL2 predicted by AlphaFold ver 2 (SRCR1 in slate, SRCR2 in purple, SRCR3, SRCR4, amine oxidase domain in lightpink) and X-ray crystal structure of Δ1-2SRCR-LOXL2 (SRCR3 in cyan, SRCR4 in green, amine oxidase domain in yellow). Zn²⁺ occupying the Cu²⁺-binding site is shown as a gray sphere. (B) 3D structure of fl-LOXL2 predicted by AlphaFold ver 2 (SRCR 1 in slate, SRCR in purple, SRCR3 in cyan, SRCR4 in green, amine oxidase domain in yellow) in two angles. The PACE4 cleavage site (³¹⁴Arg-³¹⁵Phe-³¹⁶Arg-³¹⁷Lys-↓³¹⁸Ala) is highlighted in red. The N-glycosylation sites (Asn288, Asn455, Asn644) are in orange stick. Peptides connecting domains are in gray.