Fig 1.
Ecotin orthologs from pathogenic microbes inhibit all three MASP enzymes.
Affinities of the four ecotin orthologs as well as SGMI-2 and TFMI-3 toward MASP-1 (light grey bars), MASP-2 (black bars) and MASP-3 (dark grey bars) are shown as–log10(KI) (M) values. Asterisks (*) indicate that no inhibition could be detected.
Fig 2.
KM values of Z-L-Lys-SBzl on MASP-1 (A), MASP-2 (B) and MASP-3 (C) were determined.
The KM values of MASP-1, MASP-2 and MASP-3 for Z-L-Lys-SBzl are 334μM, 618μM and 125μM, respectively. Symbols represent the average of three independent measurements. Error bars represent the SEM. The Hill equation was fitted to the data. The KM values are given in μM ± the standard error.
Table 1.
KI values of four ecotin orthologs, SGMI-2 and TFMI-3 towards all three MASP enzymes.
The ecotin-MASP affinities in Table 1 were measured at thermodynamic equilibrium as described in the Materials and methods. For tight-binding inhibition assays, enzyme and inhibitor concentrations need to be low, close to the KI value, and reaching the equilibrium requires several hours-long incubation. Then, small synthetic substrates are added, and their conversion rates report on the concentration of the free enzyme. In these short readout time assays the small substrates compete only with the relatively high-off-rate inhibitors and in this case the measured apparent KI value needs to be corrected with KM (see KM values in Fig 2). On the other hand, small substrates do not compete with the low off-rate tight-binding inhibitors for enzyme binding.
Fig 3.
E. coli ecotin inhibits the activation of pro-FD to FD in NHS.
Its efficiency is compared to that of our already published in vitro evolved MASP-3 inhibitor, TFMI-3. 1 μM ecotin increases the half-life of fluorescently labeled pro-FD in NHS from 3 hours to 10 hours, while 1 μM TFMI-3 prolongs that to ~40 hours.
Fig 4.
E. coli ecotin inhibits the MASP-3 driven pro-FD conversion.
E. coli ecotin was shown to be a high-affinity inhibitor of MASP-3 (see Table 1) in a small synthetic substrate-based in vitro enzyme-inhibition test, where enzyme and inhibitor reached thermodynamic equilibrium. We measured how efficiently ecotin inhibits MASP-3 under non-equilibrium conditions by competing with pro-FD, the natural substrate of the enzyme. We found that ecotin inhibits MASP-3 driven pro-FD conversion in a concentration-dependent manner (), but only slightly more effectively than the 20-fold lower affinity MASP-3 inhibitor TFMI-3 (
) (see Table 1). After pre-incubating ecotin with MASP-3 to approach thermodynamic equilibrium, the interaction of the two proteins became stronger and stoichiometric (
) in accordance with the subnanomolar KI value (see Table 1). This suggests that the interaction reaches equilibrium slowly. Data points are the average of two independent measurements. Error bars represent the SEM. Dotted lines are used to visually emphasize the trends.
Fig 5.
E. coli ecotin inhibits FD-driven C3bBb-formation and C3bBb-mediated C3b-production.
Approximately 45% of C3 was cleaved when no inhibitor was present (positive control) (A). When no FB and FD were present, no C3-cleavage occurred (negative control) (B). In (C) the sample contained 10 μM E. coli ecotin, while in (D) 100 μM FUT. As the AP-type C3bBb C3-convertase has a short half-life, no C3-clevage occurs after 10 minutes (E), hence, we compared the amounts of C3b generated in 10 minutes after adding FD (F). At 10 μM, E. coli ecotin decreased the production of C3b by 20% (C, F), while at 100 μM, FUT, the small-molecule general complement-inhibitor, decreased C3b-production by 80% (D, F).
Fig 6.
Ecotin is shown to inhibit LP- and AP-activation in human, and LP-activation in mouse and rat serum.
C3-deposition on mannan-coated plates from dilute sera was measured as a function of inhibitor-concentration. All four tested ecotin orthologs are equally potent inhibitors of LP-activation in 50-fold diluted NHS. Their effects are comparable to that of our previously evolved selective MASP-2 inhibitor, SGMI-2. (A) E. coli ecotin inhibits the LP in analogous tests using 60-fold diluted pooled sera of C57BL/6 and BALB/c mice. (B), and 60-fold diluted serum of Wistar rat (C). IC50 values for human and rodent serum are listed in (D) and (E), respectively. In the about 10-fold more concentrate, 6-fold diluted NHS used for the alternative pathway inhibition tests, 10 μM of the ecotin orthologs provide ortholog-specific, but significant inhibition, L. major ecotin being the most potent one (F). Symbols represent the average of at least two measurements. Error bars represent the SEM.
Fig 7.
Ecotin orthologs do not inhibit the classical pathway (CP).
CP was triggered from 50-fold diluted NHS on IgG coated plates. Symbols represent the average of four parallel measurements. Error bars represent the SD.
Fig 8.
Gating strategy for analyzing C3 and C5b9 deposition by flow cytometry.
Thresholds and gating on bacteria population on FSC-SSC plots was set up and verified by analyzing signals of 0.2 μm filtered buffer. Even the filtered buffers contained low amount of particles in the size range of bacteria (A) causing low level of inhomogeneity in the bacterial gate. A representative C3 staining of 2% NHS treated ATCC 23505 cells (filled grey histogram untreated cells, dashed line wild type cells, solid black line ecotin KO cells) is shown on panel B. Fluorescence intensity was measured on 20,000 cells captured inside the bacteria gate and C3+ marker was set on the negative control untreated sample. Because of the biasing effect of the few non-bacterial particles in the bacteria gate, the statistical analysis (MFI) of C3-deposition was conducted on the cells under the C3+ marker. Control samples had no event in this region (B). C5b9/PI positive cells were determined by double staining of the cells after treatment. 20,000 cells in the bacteria gate were analyzed, quadrants were set on the untreated, labeled samples. Sample preparation alone induces some cell death (PI positivity) in the samples. Percentage of C5b-9/PI positive cells inside bacteria gate (upper right quadrant) was used for data analysis (C).
Fig 9.
Endogenous ecotin protects E. coli against C3 and C5b-9-deposition from NHS.
The extent and the initiating pathway of CS activation were studied on two wild type and ecotin KO coli strains, the polymannan LPS carrying ATCC 23505 and the smooth LPS containing ATCC 12014. Cells were treated with increasing concentrations of NHS and labeled for deposited C3 fragments, and C5b-9 as described in Materials and methods. Mean fluorescence intensity (MFI) values of C3 positive cells (A, D) and percentage of C5b-9/PI positive cells inside the bacteria FSC-SSC gate were determined and is shown (C, F). Cells were treated with 2% NHS containing the indicated inhibitors, and the MFI of deposited C3 was determined as for panel A and D. MFI of deposited C3 without inhibitor was considered 100% complement activity (B, E). Data represent median +/- SD of three independent experiments. MFI of the C3-positive and C5b-9/PI positive ATCC 23505 ecotin KO cells was orders of magnitude higher than that of the wild type cells (A, C). This strain should trigger the LP, and indeed, SGMI-1 and SGMI-2, inhibiting MASP-1 and MASP-2, respectively, as well as exogenously added ecotin provided complete inhibition, while the anti-C1q mAb had no effect. The broad-specificity protease-inhibitor, FUT-175 was included as a control (B). Similar trends were found with the ATCC 12014 cells. This strain is less intensely attacked by the complement, and while it is also protected by its endogenous ecotin, the ratio of C3 MFI values and percentage of C5b-9/PI positive cells are smaller (D, F). All exogenously added inhibitors decreased C3-deposition on ATCC 12014 ecotin KO cells, but SGMI-1 and anti-C1q mAb were the most efficient. Therefore, this strain was attacked by all three, but dominantly by the classical and the alternative pathway (E).
Fig 10.
Intense NHS-driven cell-lysis is detected by DNA-content of cell supernatants.
We examined MAC formation both by C5b-9 specific antibodies and by propidium iodide (PI) to identify bacteria killed by MAC-driven cell lysis. In several samples, where cells were treated with above 4% NHS, the C5b-9 MFI values of the C5b-9/PI positive cells topped or even decreased (Fig 9C and 9F) indicating some artefacts. Here we show that the supernatants of these cells contain large amounts of DNA suggesting that the highest-labeled C5b-9/PI positive population has been disintegrated. Supernatants obtained during the preparation for flow cytometry experiments were applied onto agarose gel and DNA was detected with SYBR Green. No significant amounts of DNA leaked from wild type ATCC 23505 cells upon incubation with serum up to 16% NHS (A), while its ecotin KO derivative already lysed in as low as 1% NHS (B). The supernatant of ATCC 12014 cells contain detectable (C), while that of its ecotin KO derivative significant amount (D) of DNA upon treatment with 8% or 16% NHS. The % values indicate the NHS concentration the cells were treated with.
Fig 11.
Endogenous ecotin protects bacteria against bactericidal effects of serum.
Wild type cells of both strains were more resistant to the bactericidal effects of serum than their ecotin KO versions. Accordingly, when the ecotin KO cells were provided with the ecotin gene using an IPTG-inducible vector, non-induced cells behaved as the untransformed, original ecotin KO cells, while 50 μM IPTG-induction restored the resistant, wild type phenotype in these cells as follows. ATCC 23505 ecotin KO cells and their transformed but non-induced version were killed even by 1% NHS, while wild type cells and transformed, IPTG-induced ecotin KO cells were completely protected against 16% NHS (A). Up to 4% NHS, endogenous ecotin provided significant protection for the wild type and the transformed, IPTG-induced ecotin KO ATCC 12014 cells, but at higher NHS concentration, both the ecotin KO cells, regardless of IPTG-induction, and the wild type cells died (B). We also tested if heat inactivated serum (HIS) affects viability of these strains. Interestingly, above 4% concentration, HIS eliminated about 50% of the untransformed ATCC 23505 ecotin KO cells and their transformed but non-induced derivative, while the original wild type and the transformed and IPTG-induced ecotin KO version were completely protected (C). Wild type and ecotin KO ATCC 12014 cells were practically resistant to HIS, regardless of transformation and IPTG-induction (D). Results are from three parallel experiments. Error bars represent the SD, asterisks (*) represent p < 0.05 (two-tailed Student’s t-test) between the OD values of wild type and ecotin KO cells and their derivatives.
Fig 12.
Mechanism of the bactericidal effects investigated by exogenous inhibitors.
Just like for C3-deposition in Fig 9, we tested the efficacy of exogenously added inhibitors also on the bactericidal effect. We used 1% NHS where the observed bactericidal effect of heat-stable factors (see Fig 11C) is negligible. Exogenously added SGMI-1 and -2, as well as ecotin rescued the ATCC 23505 ecotin KO strain, while anti-C1q mAb exerted minimal effect (A) supporting the dominant role of LP-activation. ATCC 12014 ecotin KO cells were rescued by the anti-C1q mAb, ecotin and SGMI-1, while SGMI-2 exerted a minor effect suggesting that all three complement pathways are in action, but the classical and alternative pathways dominate (B). We also attempted to identify the factors responsible for complement-independent bactericidal effects of 16% HIS. Only ecotin and the general serine protease inhibitor FUT-175 could diminish this attack (C). Data are from at least two parallel measurements, asterisks (*) representing p < 0.05 (two-tailed Student’s t-test) when comparing the OD values of samples treated with inhibitors to those of 1% NHS. As data in (C) were obtained from two parallels, no statistical significance was calculated. Error bars represent the SD.
Fig 13.
Occurrence of ecotin orthologs in the kingdoms of Eukaryota and Bacteria.
Orthologs of ecotin are present in both Eukaryota and Bacteria kingdoms. More than three hundred species express at least one ecotin homolog. Among eukaryotic species, the Trypanosomatidae, belonging to the Kinetoplastida class, contain several species that are causative agents of serious and widespread diseases (based on publicly available pathogen databases (see Materials and methods)), and 12 species of this genus express ecotin homologs (based on the PFAM database https://pfam.xfam.org). Among Bacteria two classes are exceptionally rich in ecotin expressing species: Bacteroidetes (40 species) and Proteobacteria (158 species). Numbers in bold indicate how many pathogens exist among the ecotin expressing species in that genera; while those in italic indicate how many pathogens are among all species in that genera. Figure was made based on PFAM (https://pfam.xfam.org).