The glutaminase activity of L-asparaginase is not required for anticancer activity against ASNS-negative cells

Abstract

L-Asparaginase (L-ASP) is a key component of therapy for acute lymphoblastic leukemia. Its mechanism of action, however, is still poorly understood, in part because of its dual asparaginase and glutaminase activities. Here, we show that L-ASP's glutaminase activity is not always required for the enzyme's anticancer effect. We first used molecular dynamics simulations of the clinically standard Escherichia coli L-ASP to predict what mutated forms could be engineered to retain activity against asparagine but not glutamine. Dynamic mapping of enzyme substrate contacts identified Q59 as a promising mutagenesis target for that purpose. Saturation mutagenesis followed by enzymatic screening identified Q59L as a variant that retains asparaginase activity but shows undetectable glutaminase activity. Unlike wild-type L-ASP, Q59L is inactive against cancer cells that express measurable asparagine synthetase (ASNS). Q59L is potently active, however, against ASNS-negative cells. Those observations indicate that the glutaminase activity of L-ASP is necessary for anticancer activity against ASNS-positive cell types but not ASNS-negative cell types. Because the clinical toxicity of L-ASP is thought to stem from its glutaminase activity, these findings suggest the hypothesis that glutaminase-negative variants of L-ASP would provide larger therapeutic indices than wild-type L-ASP for ASNS-negative cancers.

Figure 1

Distinct coordinations of asparagine and glutamine in the catalytic site of E colil-ASP. Snapshots were taken at ∼20 ns of simulation. Q59 typically interacts with the backbone of both asparagine (A; light green) and glutamine (B; orange), but the patterns differ. Asparagine is usually coordinated through its backbone -NH group by the side-chain oxygen of Q59, whereas the backbone carboxyl of glutamine often interacts with the backbone -NH group of Q59, while the side chain of Q59 faces away from the substrate.

Figure 2

Enzymatic characterization of Q59 l-ASP mutants. (A) Coomassie blue–stained SDS-PAGE showing expression of l-ASP WT and Q59 mutants. The expression vector was transformed into E coli Bl-21 strain, and 20 µL of culture supernatant was analyzed by SDS-PAGE. The empty expression vector (Ctrl) and T89V (inactive mutant) served as negative controls for assays of enzyme activity in panels B and C. (B) Asparaginase activity of Q59 mutants by colorimetric assay. (C) Glutaminase activity of Q59 mutants by colorimetric assay. (D) Asparaginase-specific activity of purified Q59 mutants by the colorimetric assay. (E) Glutaminase-specific activity of purified Q59 mutants by the colorimetric assay. (F) Ratio of glutaminase- and asparaginase-specific activities for purified l-ASP mutants. SDS-PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

Figure 3

Asparagine and glutamine deamidation kinetics. WT l-ASP or Q59L l-ASP was added to a reaction solution containing 100 µM asparagine, 1600 µM glutamine, or both. Concentrations of asparagine (A) and glutamic acid (B) were measured over a 1500-s time series by LC-MS/MS. Solid symbols represent concentrations in the single substrate reaction, and open symbols represent concentrations in the mixture.

Figure 4

Anticancer activity of WT, Q59L, and Q59F l-ASP. (A-H) Two ovarian cancer cell lines (OVCAR-8 and SK-OV-3) and 6 leukemia cell lines (MOLT-4, K562, NALM-6, REH, SR, and CCRF-CEM) were seeded in 96-well plates, incubated for 48 hours, treated with a range of (WT, Q59L, or Q59F) l-ASP concentrations for 48 hours, and finally assayed with CellTiter-Blue using fluorescence excitation at 544 nm and emission at 590 nm. (I-J) MOLT-4 and OVCAR-8 cells were seeded in 96-well plates and incubated for 48 hours, then treated with indicated concentrations of E coli l-ASP WT or Q59 mutant for 48 hours. Inhibition of cell viability was measured as in panels A-H. Sham treatment was used as a control. (K) Western blot analysis of ASNS levels in the indicated cells treated with an EC₅₀ dose of l-ASP. (L) Western blot analysis of ASNS levels in OVCAR-8 cells treated with l-ASP mutants. Numbers below the blot represent the relative level of ASNS, which was normalized to the level of the loading control β-actin (set to “1” for the control).

Figure 5

Selective growth inhibition of ASNS-negative cancer cells by WT, Q59L, and Q59F l-ASP. (A) Western blot analysis of ASNS levels in OVCAR-8 cells after 48 hours transfection with ASNS siRNA (siASNS) or negative control siRNA (siNeg). β-actin was used as a loading control. (B) WT and Q59L l-ASP concentration-activity curves. The OVCAR-8 cell line was transfected with negative control siRNA (siNeg) or ASNS siRNA (siASNS) for 48 hours, then treated with a range of l-ASP concentrations for 48 hours, and finally assayed with CellTiter-Blue. WT, Q59L, and Q59F l-ASP concentration-activity curves were determined in the (C) Sup-B15 and (D) RS4;11 leukemia cell lines by CellTiter-Blue assay. (E) Western blot analysis of ASNS levels in Sup-B15 and RS4;11 cells treated with an EC₅₀ dose of l-ASP. No treatment was used as a primary control, and MOLT-4 cells were included as a secondary control.

Figure 6

Proposed model for the mechanism of WT l-ASP’s and Q59L lASP’s anticancer activity. The mechanism of anticancer activity depends on l-ASP glutaminase activity and ASNS expression, which is reflected by the color gradient of the background. For simplicity, glutamine synthesis pathways are not shown. (A) (Left panel) (1) Q59L l-ASP effectively depletes Asn but not Gln, which (2) is imported by the cancer cell for (3) synthesis of Asn by ASNS, thereby promoting cancer cell proliferation (4). Numbering is omitted from subsequent panels, but analogous interpretation illustrates that the added glutaminase activity of WT l-ASP decreases the extracellular supply of Gln, thereby limiting cancer cell proliferation (right panel). (B) Low-ASNS cancer cells are insensitive to Q59L l-ASP (left panel), but not to WT l-ASP (right panel). (C) ASNS-negative cancer cells are sensitive to both Q59L (left panel) and WT (right panel). Details of the model are provided in the text.