Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Macrolides, including erythromycin (Er), are antibiotics with bactericidal effects. There is, however, an increasing body of evidence demonstrating that macrolides may have activities not related to their antibiotic properties. Er A and its derivatives (especially 14-member-ring macrolides, such as roxithromycin and clarithromycin) can significantly inhibit the neutrophil oxidative burst in vitro (1). The same compounds have also been shown to reduce spontaneous interleukin (IL)-6 (2) and IL-8 expression in human bronchial epithelial cells (3) and the Hemophilus influenzae endotoxin-induced release of IL-6 and IL-8 by these cells (4). In addition, macrolides are able to inhibit lipopolysaccharide (LPS)-induced goblet-cell hypersecretion in an animal model (4). Suppression of LPS-induced intrapulmonary neutrophil influx (5), and activities on airway epithelial cell and gland electrolyte secretion (6, 7), may both contribute to this inhibition. Taken together, such a variety of anti- inflammatory effects account for the in vivo efficacy of macrolides in conditions such as diffuse panbronchiolitis, cystic fibrosis, and bronchiectasis characterized by a "vicious circle" leading to mucus hypersecretion (8).

Human neutrophil elastase (HNE) is an important component in the establishment of the vicious circle in the previously mentioned conditions (13). It has been reported that Er treatment results in a reduction of neutrophil-derived elastolytic activity in the airways of subjects with chronic airway disease (8) and diffuse panbronchiolitis (14). Such a feature was, however, attributed to a direct effect of Er on neutrophil function. The structure of Er is characterized by a macrocyclic lactone ring (Figure 1). Interestingly, several types of lactones (ynenol, protio, and halo enol lactones) are heterocyclic compounds capable of inhibiting HNE (15). Such structural homology raises the question of whether Er might exert a direct inhibitory effect on HNE.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Molecular structure of Er and FE A. Inset: molecular structure of S(L) -aryl lactone (20).

We therefore designed a series of experiments aimed at evaluating whether or not Er is an HNE inhibitor and, if so, by what mechanism this activity is exerted. In this study, experiments were also performed with the semisynthetic macrolide flurythromycin (FE) base, a 14-membered 8-fluoro-macrolide (Figure 1) obtained by mutasynthesis using a mutant strain of the Er producing microorganism (16), to understand whether 14-membered fluorinated macrolides retain the same activity and mechanism of action as Er.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals and Equipment

In this study we used phosphate buffer 0.1 M NaH₂PO₄- Na₂HPO₄, pH 8.0, with 0.05% wt/wt Triton X-100 and 0.5 M NaCl. HNE was purchased from ART (Athens, GA) and active-site titrated in all experiments by MeO-Suc-Ala-Ala-Pro-Val-CH₂-Cl, as described elsewhere (17). HNE was dissolved in 50 mM CH₃COONa, pH 5.5, supplied with 150 mM NaCl and diluted in the previously described phosphate buffer. Chromogenic substrate MeO-Suc-Ala-Ala-Pro-Val-pNa (Sigma, St. Louis, MO) was dissolved and diluted in Me₂SO. Er was purchased from Sigma, and FE base (Pierrel SpA, batch #60296) was a kind gift from Pharmacia & Upjohn (Milan, Italy); both were dissolved and diluted in Me₂SO. Proflavine (Sigma) was dissolved and diluted in phosphate buffer as described earlier. Hydrazine (Sigma) was diluted in distilled water. Assays were performed at 26 ± 1°C using a Beckman DU 650 L spectrophotometer. All assay cuvettes contained enough Me₂SO to obtain a final cosolvent concentration of 10%. Changes in absorbance at 410 and 466 nm were recorded in cuvettes with 1 cm optical course and through a total volume of 1 ml.

Definition of Michaelis Constant

A constant amount of HNE (2.6 × 10⁸ M) and serial dilutions of substrate (ranging from 2 × 10³ M to 9.8 × 10⁷ M) were combined into the cuvettes. Reactions, recorded in triplicate over a time interval of 3 min, were started with addition of substrate dilutions. Estimations of maximal velocity (V_max) and Michaelis constant (K_m) were made according to Eisenthal and Cornish-Bowden (18).

K_m (± standard deviation [SD]), 4.2685 ± 0.1236 × 10⁵ M.

catalytic rate constant (k_cat) (± SD), 4.131 ± 0.5986 s¹.

k_cat/K_m, 96,778.7 M¹ s¹.

V_max (± SD), 0.1325 ± 0.0159 s¹.

Enzyme Inhibition Assay Method A: No Deacylation Delay (19)

The concentrations used for enzyme inactivation were: HNE, 1 × 10⁶ M; inhibitor(s) (Er or FE), 5 × 10⁵ M; and substrate 3 × 10⁴ M. The inhibitor(s) solution was added to a buffered enzyme solution and, at various intervals over the course of 2 h, aliquots were withdrawn and diluted 50 to 100-fold into a cuvette containing a buffered solution of substrate. Substrate hydrolysis was read immediately at 410 nm and was used to determine the remaining enzyme activity versus a control solution without inhibitor. A semilogarithmic plot of enzyme activity (ln E/E₀) versus time showed the inhibition rate k_obs that was obtained from the slope of the initial linear portion of the curves.

Enzyme Inhibition Assay Method B: 10-Min Deacylation Delay (19)

This method is identical to method A, with the exception that aliquots were diluted into cuvettes containing buffer. After 10 min, during which unstable acyl enzyme intermediates were allowed to decompose, substrate was added and the rates were measured at 410 nm. At the end of the 2 h, the enzyme incubation solutions showing greater than 10% inhibition with respect to the control were treated with 50 mM hydrazine and enzyme activity was monitored for additional 30 min.

Competitive Substrate Assay: Determination of the Substrate Binding Constant (K_s) or Inhibition Constant (K_i), and Acylation Rate Constant (k_a) (19, 20)

Concentrations used for the assays were: HNE, 5 × 10⁹ M; substrate, 3 × 10⁵ M; Er and FE ranging from 1 × 10⁷ M to 5 × 10⁵ M. Reactions started with the addition of HNE and were recorded in triplicate over 30 min. During this assay it was assumed that concentrations of both substrate and inhibitor remained constant, due to the large excess of those with respect to the enzyme. When inhibitors were incubated with HNE and substrate there was competition between inhibitor and substrate in occupying the HNE active site. Kinetic analysis, according to Main (21), was derived from competitive assay between substrate and a time- dependent, irreversible inhibitor with corrections, according to Baek and colleagues (20), due to inhibitors that are alternate substrates. The limiting slope of the progress curves, due to deacylation of the inhibitor acyl enzyme, was estimated and subtraction of this rate from the curves allowed accurate determination of the true exponential approach to steady-state (k_obs). The absorbance changes against time were corrected for the limiting slope and the semilogarithmic plots. These results gave a straight line with a slope of k_obs (first-order rate constant).

Main (21) described the progress curves for substrate consumption in time-dependent inhibition with equations (1) and (2):

(1)

(2)

where S₀ and I₀ are the initial concentrations of substrate and inhibitor, respectively; and V₀ and V are the velocity of substrate hydrolysis at initial time and time t, respectively.

Rearrangement of the definition of k_obs gives equation (2). A plot of I₀/k_obs (second-order rate constant) versus the initial inhibitor concentration I₀ gave a straight line with the slope 1/k_a and x intercept of K_s(1 + S₀/K_m) or K_i(1 + S₀/K_m).

In our assays condition S₀ was 30 µM and K_m was 42,685 µM.

Determination of Deacylation Rate Constant (k_d) for Macrolide Inhibition: Proflavine Displacement Assay (20)

Maximal difference in absorbance between free proflavine and enzyme-proflavine complex is recorded at 466 nm, so the measurement of absorbance increase at this wavelength can be used to monitor acyl-enzyme decay. When an inhibitor (3.3 µM) was added to a solution of HNE (3.3 µM) and proflavine (0.33 µM) and the time-dependent absorbance change was monitored at 466 nm for 15 min (in triplicate), an initial fast disappearance of the enzyme-dye spectrum was seen. The decrease in absorbance was due to enzyme acylation and thus dissociation of proflavine, whereas the slow increase was due to deacylation of acyl-enzyme and rebinding of proflavine to free enzyme. The rate constant of deacylation of both mixtures and of separated enantiomers (see RESULTS) was calculated directly from the second phase of the progress curve of the enzyme-dye spectrum (method A, according to Baek and associates [20]).

Ultimate Activity Assays (19, 22)

A series of incubation solutions containing inhibitor(s) to enzyme molar ratios ranging from 0 to 50 were prepared mixing HNE, 1 × 10⁶ M, and a solution of inhibitor ranging from 0 to 5 × 10⁵ M. The enzyme activity remaining 16 h after the addition of the inhibitor was determined by diluting 1:20 an aliquot of the incubation mixture into the buffer. After 10 min, the substrate (3 × 10⁴ M) was added and substrate hydrolysis was monitored at 410 nm. A plot of the final enzyme activities versus the initial inhibitor-to-enzyme concentration ratios gave a straight line with an x intercept representing the number of turnovers per inactivation.

Hydrazine Reactivation Assays (19, 22)

Inactivated enzyme solutions were treated with the nucleophile hydrazine (5 × 10² M), and enzyme activity was monitored over 40 h. HNE (1 × 10⁶ M) was totally inhibited with a 50-fold excess of inhibitor(s). After 60 min, the aqueous hydrazine solution was added and enzyme activity versus a control solution (no inhibitor) was monitored over the course of a 40-h period. This was accomplished by periodic 1:20 dilutions of aliquots of incubation solutions into the buffer. After 10 min, the remaining enzyme activity was assayed with 3 × 10⁴ M substrate. Reactivated enzyme activity was plotted versus time.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Inhibition Assays

Er at 50 µM rapidly (30 min) inhibited 15% of HNE activity and the inactivation rate progressively decreased, until cessation. However, inhibition was temporary and full activity was regained within 1 h. Temporary inhibition of this nature is consistent with the formation of a moderately stable acyl enzyme intermediate from Er and HNE that causes a temporary suppression of HNE activity until the inhibitor is completely turned over (19). Recovery of enzyme activity was rapid (15 min) after addition of the deacylating nucleophile hydrazine (50 mM). Er acts as a potential alternate substrate inhibitor and so should interact with the enzyme in the manner of a normal substrate.

The kinetic model shown in equation (3) describes the transient inhibitory activity of an alternate substrate inhibitor (19, 20):

(3)

where I is the inhibitor, E.I the Michaelis complex, E*I the acyl enzyme (hydrazine-susceptible), and I' the hydrolyzed inhibitor.

FE (50 µM) showed a significant inhibition of HNE that was time-dependent and began immediately upon mixing. The recovery of enzyme activity observed within 30 min after the addition of 50 mM hydrazine was < 10%. We supposed that the complete inhibition was the result of either direct alkylation of the acyl enzyme intermediate by the reactive inhibitor moiety, or the formation of an extremely stable acyl enzyme intermediate. FE acts as an inactivator, with a mechanism of inactivation described in equation (4) (19):

(4)

where I is the inactivator, E.I the Michaelis complex, E*I the acyl-enzyme, E  I the permanently inactivated enzyme (hydrazine-resistant), and E + I' the free enzyme.

Acyl enzyme formation proceeds as with a normal substrate in a manner that is consistent with a mechanismbased inhibition process which can be described by K_i, k_a, but it is followed by an inactivation event (alkylation) k₂ which competes with deacylation k_d.

Acylation of HNE by Er and FE: Competitive Substrate Assay

The rate of substrate turnover, measured continuously during the course of the enzyme acylation by the inhibitor(s), decreased at a first-order rate constant, k_obs (Figure 2). K_s or K_i characterizes the affinity of the inhibitor toward the active site of HNE and is a measure of the stability of the E.I complex (20). In both cases, K_s or K_i was about 10⁶ M (Figure 3A); no great differences were observed in k_a and k_d (Figure 3B). The second-order rate constants for the two inhibitors, derived from the competitive substrate assay (Figures 2B and 2C) and confirmed by the inhibition assay without deacylation delay (Figure 2A), were good. According to Reed and Katzenellenbogen for protio and/or halo lactones (19), the observed specificity of the inhibitors (Table 1) was consistent with an acyl-enzyme structure similar to that adopted by a peptide substrate (23, 24).

View larger version (18K): [in this window] [in a new window]   Figure 2.   (A) Enzyme inhibition assay method A (no deacylation delay). Inhibitor solution (50 µM) was added to 1 µM HNE solution. At different times, aliquots were withdrawn and diluted into a 300-µM substrate solution. Remaining enzyme activities versus control solution (no inhibitor) were plotted versus time. The slope of the initial linear portion of the curves shows the inhibition rate kobs. (B and C) Inhibition progress curves (competitive substrate assay). Solutions of inhibitor ranging from 0.1 to 50 µM substrate were started with addition of 5 nM HNE and assays were recorded over time. The optical density changes against time gave the first-order rate constant kobs (slope).

View larger version (10K): [in this window] [in a new window]   Figure 3.   (A) Determination of Ks or Ki using a competitive substrate assay. Assay conditions were as described in Figures 2B and 2C. The second-order rate constants I0/kobs were plotted versus the initial inhibitor concentrations, giving a straight line with a slope of 1/ka and an x intercept. This allowed us to calculate Ks for Er and Ki for FE as described in MATERIALS AND METHODS. (B) Measurement of kd by the proflavine displacement assay. Time-dependent absorbances were monitored over time. Deacylation rate constants were calculated directly from the second phase of the progress curves.

Deacylation of HNE by Er and FE: Proflavine Displacement Assay

Deacylation of both macrolides resulted in a biphasic curve (Figure 3B). This occurrence was probably due to the presence in Er and FE solutions of two enantiomers, of which one was in large excess over the other, as confirmed by capillary electrophoresis experiments (data not shown). Compared with k_a, k_d is very slow; this step is therefore truly the rate-determining step for Er but not for FE, where k₂ is the most probable pathway (Table 2).

Efficiency of the Inactivation of HNE by Er and FE: Ultimate Activity Assay

The inactivation efficiency for inhibitors of HNE was determined by measurement of the partition ratios or the number of turnovers per inactivation (number of turnovers per inactivation = partition ratio + 1) (22) (Figure 4 and Table 3). At low inhibitor-to-enzyme concentration ratios, inhibitor turnover caused significant deviation from linearity and incomplete inhibition. FE acylates the enzyme and the acyl enzyme will partition between being inactivated and deacylated; Er, on the other hand, acylates the enzyme, which deacylates. The number of equivalent of inhibitor necessary to inactivate the enzyme completely was determined for a period to allow complete consumption of inhibitor(s) (19). HNE was inactivated more efficiently by FE than by Er.

View larger version (8K): [in this window] [in a new window]   Figure 4.   Efficiency of HNE inactivation by the inhibitors (ultimate activity assay). A series of incubation solutions containing inhibitors:enzyme molar ratios ranging from 0 to 50 were prepared mixing 1 µM HNE and a solution of inhibitor ranging from 0 to 50 µM. Remaining enzyme activities, 16 h after the addition of the inhibitor, were determined by first diluting aliquots of the incubation mixtures into the buffer and then adding 300 µM substrate. The plot of the final enzyme activities versus initial inhibitor:enzyme concentration ratios gave a straight line. The x intercept represents the number of turnovers per inactivation.

Permanence of Inactivation of HNE by FE: Hydrazine Reactivation Assay

The nature of the covalent attachment between the inhibitors and HNE in the completely inactivated enzyme species was investigated by treatment with the nucleophile hydrazine for an extended period of time. The relatively labile acyl serine linkage of a stable acyl enzyme is generally cleaved rapidly by hydrazine, whereas covalent attachment via alkylation by the reactive moiety on the inhibitor(s) is stable (19). When HNE that had been completely inactivated by FE was treated with 50 mM hydrazine and monitored for enzyme activity, less than 10% of control activity was recovered over a period of 1 h (Figure 5 and Table 3). This shows that haloketone, tethered to the active-site serine, alkylates an active-site nucleophile to give an inactivated enzyme. By contrast, when HNE that had been completely inactivated by Er was treated with hydrazine and monitored over 40 h, a recovery of about 50% was observed within 5 h (> 80% at 30 h) (Figure 5, inset; and Table 3). The inactivated enzyme possesses a less-stable covalent attachment that is susceptible to cleavage by hydrazine.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Hydrazine reactivation of HNE (40 h) inhibited by Er or FE. A solution of 1 µM HNE was totally inhibited by 50 µM inhibitor. After 60 min, 50 mM hydrazine solution was added, and aliquots of the incubation solution were periodically diluted in the buffer. Plotted aliquots were then tested with 300 µM substrate. Reactivated enzyme activities were plotted versus time. Inset: magnification of the first 2-h analysis.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this paper we provide evidence that Er and FE, two 14-member-ring macrolides, can inhibit HNE.

The presence of a macrocyclic lactone ring in such compounds prompted us to explore the possibility that the macrolides investigated act as HNE inhibitors in a fashion similar to that displayed by lactones. Such heterocyclic elastase inhibitors are capable of being attacked by the nucleophilic active serine hydroxyl of the proteinase to give a ring-opened acyl enzyme intermediate. On the basis of the structure of the acyl enzyme intermediate, they may be categorized into two groups (19). Protio enol lactones act simply as substrates, forming very stable acyl enzyme intermediates, and the enzymes remain inhibited for the lifetime of the acyl enzyme species. Protio enol lactones belong to the group of alternate substrate inhibitors (or simple acylating agents), which also comprises benzoxazinones, 3-alkoxy-4-chloroisocoumarins, isobenzofuranones, and saccharin derivatives (15). By contrast, halo enol lactones and ynenol lactones are irreversible inhibitors with a latent, reactive functional group that is unmasked upon reaction with the active-site residue of the enzyme. This results in an acyl enzyme intermediate consisting of an enzyme tethered to a reactive functional group at the active site (Ser-195), and a simultaneous unmasking of the reactive group, giving a covalently modified enzyme that is permanently inactivated (15, 19). Such lactones are categorized as inactivators (also called "suicide inhibitors," k_cat inhibitors, or enzyme-activated inhibitors), and the category also includes cephalosporins, 3,4-dichloroisocoumarins, 3-alkoxy- 7-amino-4-chloroisocoumarins, and halomethyldihydrocoumarins (25).

Similar to lactones, macrolides studied in this work have an -O-CO-group. The -CH₃ group in  position (Figure 1) allows us to consider them as S enantiomers (19). This stereochemistry makes Er and FE good HNE substrates (20) (Figure 1, inset). Er acts as an alternate substrate inhibitor, its action appearing to involve acyl transfer to Ser-195, with generation of an acyl enzyme (26), according to the scheme reported in Figure 6. By contrast, FE acts as an inactivator: enzyme acylation results in the formation of a halo ketone derivative which can alkylate an active site nucleophile (probably His-57) to give an inactivated enzyme (26). An alternative scheme for the putative inactivation mechanism of HNE by FE is summarized in Figure 7: (1) HNE Ser-195 attacks the "lactonic carbon" to generate an acyl enzyme; (2) HNE His-57 catalyzes a trans elimination of the carbohydrate moiety; and (3) a nucleophilic attack to the acrylic ester structure, most likely catalyzed by HNE His-57, generates a stable doubly covalent adduct. Due to stereochemical difference in Er, the previously cited trans elimination reaction would not be possible.

View larger version (20K): [in this window] [in a new window]   Figure 6.   Scheme for HNE inhibition by Er as alternate substrate inhibitor. Generation of an acyl enzyme by attack of Ser-195 at the lactone bond of Er leads to an acyl enzyme which undergoes slow deacylation.

View larger version (23K): [in this window] [in a new window]   Figure 7.   Scheme for HNE inhibition by FE as an irreversible inactivator. An acyl enzyme is generated by HNE Ser-195 attack at the lactone carbonyl of FE. HNE His-57 catalyzes trans elimination of the carbohydrate moiety. A nucleophilic attack by His-57 at the acrylic ester structure generates a stable doubly covalent adduct.

Both Er and FE are good HNE inhibitors because they have a high capacity of acylation (k_a) and a low velocity of deacylation (k_d). Because the K_s (or K_i) ratio FE: Er  2, in the case of FE, the E  I interaction in the Michaelis complex is poorer than that between Er and HNE. Because the number of turnovers per inactivation of Er is  20-fold higher than that of FE (Table 3), we hypothesized that the inactivated enzyme (E  I) HNE-FE was more stable than HNE-Er. Results of the hydrazine reactivation of the acyl enzyme (Figure 5) showed that the acyl enzyme (E*I) HNE- FE was more stable than the acyl complex HNE-Er. For Er we identified only a deacylation constant (k_d), whereas for FE, in addition to the k_d we were able to calculate an alkylation constant (k₂) correlated to a fully inactivated enzyme. Rates of enzymatic deacylation were identical for all the substrates. However, for FE the determining step was correlated to k₂ connected to a reaction with 20-fold higher rate with respect to k_d. Therefore, FE is able to inactivate HNE almost fully. However, plots of hydrazine reactivation (Figure 5, inset) show that the percentages of reactivated enzyme at 30 min were 28.5 and 7% for Er and FE, respectively, both in comparison with control. We then hypothesized that, at most, 7% of enzyme was exclusively affected by acylation, whereas the remainder was affected by both acylation and alkylation. The same plots at 40 h showed that the percentages of reactivated enzyme were 83 and 0.11% for Er and FE, respectively (Figure 5).

We can therefore conclude from our kinetics data that Er acts as an alternate substrate HNE inhibitor, whereas FE acts as an HNE inactivator. However, only crystal structure data would allow us to advance a more detailed hypothesis to elucidate inhibition mechanisms and enantioselectivity of these macrolides.

It remains to be established whether or not the inhibitory effect of the two 14-member-ring macrolides shown in this paper may in part account for the effects displayed in vivo by macrolides in conditions characterized by an excess of neutrophil elastase within the airways. Animal models of acute and chronic lung injury due to HNE would help, at least in part, to address such an issue. On the basis of the K_s or K_i value obtained, we may suppose that both Er and FE have moderate inhibitory activity in vivo and, on the basis of the inhibition mechanism, that FE would be more effective than Er. However, if our kinetic findings are relevant in vivo, inhibition of HNE might be an additional explanation for the reported effect of macrolides of reducing mucus hypersecretion in chronic conditions characterized by an excess of HNE within the airways, such as cystic fibrosis and disseminated bronchiectasis (11, 12). In fact, it is well known that HNE plays a crucial role in such conditions, causing direct stimulation of mucus secretion from airway glands (27) and stimulation of IL-8 production from epithelial cells (28) which, being a potent neutrophil chemoattractant, perpetuates a further HNE release (29). Macrolides have been reported to be able to suppress these events in vitro (3). Thus, a putative in vivo inhibition of HNE by macrolides might block two important mechanisms of the vicious circle leading to mucus hypersecretion, thus providing a further explanation for the more-than- expected efficacy of macrolides in this context. In addition, it would be interesting to investigate in future studies second-generation macrolide antibiotics, such as azithromycin or clarithromycin, and, maybe more interestingly, novel organic compounds produced by genetic manipulation (30) under this point of view.