SP-A binds alpha1-antitrypsin in vitro and reduces the association rate constant for neutrophil elastase.

BACKGROUND: Alpha1-antitrypsin and surfactant protein-A (SP-A) are major lung defense proteins. With the hypothesis that SP-A could bind alpha1-antitrypsin, we designed a series of in vitro experiments aimed at investigating the nature and consequences of such an interaction. METHODS AND
RESULTS: At an alpha1-antitrypsin:SP-A molar ratio of 1:1, the interaction resulted in a calcium-dependent decrease of 84.6% in the association rate constant of alpha1-antitrypsin for neutrophil elastase. The findings were similar when SP-A was coupled with the Z variant of alpha1-antitrypsin. The carbohydrate recognition domain of SP-A appeared to be a major determinant of the interaction, by recognizing alpha1-antitrypsin carbohydrate chains. However, binding of SP-A carbohydrate chains to the alpha1-antitrypsin amino acid backbone and interaction between carbohydrates of both proteins are also possible. Gel filtration chromatography and turnover per inactivation experiments indicated that one part of SP-A binds several molar parts of alpha1-antitrypsin.
CONCLUSION: We conclude that the binding of SP-A to alpha1-antitrypsin results in a decrease of the inhibition of neutrophil elastase. This interaction could have potential implications in the physiologic regulation of alpha1-antitrypsin activity, in the pathogenesis of pulmonary emphysema, and in the defense against infectious agents.

Background

Alpha1-antitrypsin (α1-AT) and surfactant protein-A (SP-A) are major defense glycoproteins in the alveolar spaces of human lungs. α1-AT, a 52,000 D glycoprotein, is secreted mostly by hepatocytes, and, to a lesser extent, by lung epithelial cells and phagocytes. α1-AT inhibits a variety of serine proteinases by its active site (Met358-Ser359), but its preferential target is human neutrophil elastase (HNE) as demonstrated by the high association rate constant (Kass) for this proteinase [1]. In the lungs, α1-AT protects the connective tissue from HNE released by triggered neutrophils; as a result, subjects homozygous for the common deficiency variant Z α1-AT (associated with 15% of normal plasma α1-AT levels) develop pulmonary emphysema early in life, especially if they smoke [2].

SP-A, a member of the collectin (collagen-lectin) family [3], is one of the proteins of surfactant. Structurally, it comprises an N-terminal collagen-like domain connected by a neck to a C-terminal carbohydrate recognition domain (CRD) [4]. Six trimers are linked by disulfide bridges in an octadecamer of 650,000 D, in a "flower bouquet" alignment pattern [4,5]. A complex, predominantly triantennary, carbohydrate chain of ~4,000 D [6] is attached to the asparagine at position 187 of the CRD [7]. SP-A is mainly present in the alveoli in association with phospholipids, only 1% being present in the free form [8,9]. The primary function of surfactant is to reduce alveolar surface tension at end expiration. It is now however clear that SP-A, together with SP-D, another hydrophilic surfactant protein, plays a major role in the innate defenses of lung [5-10]. SP-A, in particular, is able to bind several micro-organisms and enhance their uptake by phagocytes, stimulate the production of free oxygen radicals, and induce phagocyte chemotaxis [11].

Most binding to micro-organisms, including influenza and herpes simplex viruses, Gram-positive and Gram-negative bacteria, mycobacteria, fungi, and Pneumocystis carinii, occurs via the CRD and is inhibited by sugars or calcium chelators [12].

Since some SP-A is present in the alveoli in the free form, it has a chance of coming into contact with α1-AT. We hypothesized that, in analogy with what happens with infectious agents, SP-A could bind to α1-AT, which carries 3 biantennary or triantennary asparagine-linked carbohydrate chains [13].

In this paper we provide in vitro evidence that the inhibitory activity of α1-AT towards HNE is significantly decreased in the presence of SP-A, probably as a consequence of SP-A binding to α1-AT. Such an interaction would represent a novel mechanism of regulating alveolar α1-AT. This could have relevance both for the pathogenesis of emphysema in patients with the Z α1-AT variant and for the lungs' defenses against infectious agents.

Methods

Preparative procedures

All reagents were of analytical grade, unless otherwise specified. The buffer used in all experiments was 0.2 M Na-K phosphate, with 0.5 M NaCl, 2 mM CaCl2, and 0.05% w/w Triton × 100, pH 8.0 (phosphate buffer), unless otherwise specified. Lipopolysaccharide (LPS) from E. coli serotype 026:B6 (Sigma) and methyl-α-D-mannopyranoside (MNOCH3) (Sigma) were dissolved in phosphate buffer. HNE and human α chymotrypsin (αChy) (ART, Athens, GA) were dissolved in 50 mM sodium acetate, 150 mM NaCl, pH 5.5 and diluted with phosphate buffer. N-glycocosidase F from Flavobacterium meningosepticum (PNGase F; EC 3.5.1.52) was purchased from Roche Diagnostics (Monza, Italy). Clostridium histolyticum collagenase type III (EC 3.4.24) came from Calbiochem (La Jolla, CA). The chromogenic substrates MeOSucAlaAlaProValNA (for HNE) and SucAlaAlaProPheNA (for αChy), and the irreversible inhibitors MeOSucAlaAlaProValCMK (for HNE) and TosPheCMK (for αChy) (all from Sigma) were dissolved in (CH3)2SO. Wild-type α1-antitryspin (M α1-AT) was either from ART or purified from human serum by covalent chromatography. Capillary isoelectric focusing (CIEF) with bare fused-silica capillaries filled with polyethylene oxide and carrier ampholyte solutions in the pH 3.5–5.0 range [14] was applied to confirm the presence of the common M α1-AT variant. Z α1-AT variant was purified by covalent chromatography from subjects identified within the Italian screening program for α1-AT deficiency [15]. SP-A was isolated as described [16] from surfactant obtained from 3 patients affected by pulmonary alveolar proteinosis (PAP), subjected to therapeutic whole lung lavage [17] and from adult New Zealand rabbits. To isolate surfactant, the bronchoalveolar lavage fluid was filtered through gauze and centrifuged at 150 g for 10 minutes. The supernatant was centrifuged for 30 minutes at 80,000 × g and the resulting pellet was suspended in 10 mM Tris-HCl pH 7.4, 145 mM NaCl, 1.25 mM CaCl2, 1 mM MgCl2, 2.2 M sucrose (solution A), overlaid with 10 mM Tris-HCl pH 7.4, 145 mM NaCl, 1.25 mM CaCl2, 1 mM MgCl2, 2 M sucrose (solution B) and ultracentrifuged overnight at 85,000 × g in a Ti 60 rotor (Beckman). The floating material was dispersed in water and centrifuged for 30 minutes at 100,000 × g and the pellet recovered was stored at -70°C (purified surfactant). To obtain SP-A, surfactant was injected into a 50-fold excess by volume of 1-butanol and stirred at room temperature for 30 minutes. After centrifugation, the pellet was suspended in 1-butanol and re-centrifuged at 4,000 × g for 1 hour at room temperature. The final precipitate was dried under nitrogen and then resuspended in 5 mM Tris-HCl, 145 mM NaCl, 20 mM octyl β-D-glucopyranoside, pH 7.4 (solution C). After centrifugation at 100,000 × g for 1 hour, the pellet was resuspended in 5 mM Tris-HCl pH 7.4 (solution D) and dialyzed against solution D for 48 hours with at least six changes. The final solution was centrifuged at 100,000 × g for 1 hour and the resulting supernatant, containing purified SP-A, was stored. Endotoxin-free SP-A was obtained by treatment with polymyxin-B (Sigma). Small aliquots of SP-A in solution D were incubated in a 1:1 ratio for 6 hours at 4°C with polymyxin-agarose previously equilibrated with 5 mM Tris-HCl, 100 mM octyl β-D-glucopyranoside and 2 mM EDTA, pH 7.4. Polymyxin-agarose was removed by centrifugation at 14,000 × g for 15 minutes, and the supernatant was then dialyzed against 5 mM Tris-HCl pH 7.4 for 48 hours with at least six changes and lyophilized [17,18]. For some experiments polymyxin-treated SP-A was further purified by D-mannose sepharose 4B chromatography. SP-A was added to a small column containing D-mannose sepharose 4B (Pharmacia) previously equilibrated with 5 mM HEPES, 0.4% Triton × 100 and 1.5 mM CaCl2, pH 7.2 (solution E), and the column was washed extensively with solution E. SP-A was finally eluted with 5 mM HEPES, 0.4% Triton × 100 and 2.5 mM EDTA, pH 7.2 (solution F).

Modification of the native proteins

Native and modified proteins used in our experiments were at high degree of purification (Figure 1). See additional file 1for more details.

Figure 1

SDS-PAGE under reducing conditions. A: α1-AT; B: SP-A. Lane 1: molecular weights; lane 2: native protein; lane 3: deglycosylated protein. The two bands in gel B, lanes 2 and 3 correspond to dimers (57 and 50 kDa, respectively) and monomers (33 and 27 kDa, respectively) of SP-A.

Identification of the SP-A/ α1-AT complex

1) Gel filtration HPLC

A mixture of SP-A (1.62 mg/ml) and α1-AT (1 mg/ml) in a 1:50 molar ratio was incubated for 24 hrs at 37°C in phosphate buffer. The SP-A/α1-AT mixture and single proteins were loaded in a Jasco PU 980 HPLC system (Japan Spectroscopic, Tokyo, Japan) equipped with two Biosep-SEC-S 4000 columns (300 × 7.80 mm each, Phenomenex, Torrence, CA, USA) connected in series. Samples were eluted with 100 mM Na2HPO4, 2 mM CaCl2, pH 6.8 at a flow rate of 0.3 ml/ min, and monitored at 220 nm. The excluded (V0 = 12.43 ml) and total (Vt = 24.82 ml) volumes were determined using dextran and creatinine, respectively; a calibration curve was obtained by running through the column a set of standard proteins: α2-macroglobulin (725 kD), aldolase (158 kD), bovine serum albumin (67 kD), chymotrypsinogen (25 kD), and cytochrome C (12.5 kD). The results were reported as mean ± SD of three separate experiments. 2

2) Qualitative immunodetection by ELISA

250 ng of standard α1-AT, purified SP-A, and SP-A/α1-AT complex collected from the Size Exclusion Chromatography experiments, were immobilized in 50 mM Na2CO3, pH 9.5 overnight at 4°C in a polypropylene plate (Corning, New York, USA). Plates were then brought at room temperature, washed with 150 mM NaCl, 0.1% Tween 20 (ELISA buffer), blocked for 1 h with 50 mM Na2CO3, 2% BSA pH 9.5, incubated for 2 hrs in the presence of primary antibodies diluted 1:500 (goat anti-human α1-AT and rabbit anti-human SP-A; ICN, Aurora, OH, USA), washed and finally reacted for 2 hrs with the appropriate biotinylated secondary antibodies diluted 1:5000 (Chemicon, Temecula, CA, USA). After washing, 100 μL of avidin diluted 1:2000 were added, and samples were incubated for 30 min. Color development was achieved by incubating the samples with 1,2-phenylenediamine dihydrochloride (Dako, Bucks, UK). The reaction was stopped by addition of 100 μl of 0.5 M H2SO4 and OD was read at 490 nm with a Bio-Rad 680 Microplate Reader (Bio-Rad Laboratories, CA, USA).

Kinetics studies

Rate constants were derived by competition experiments of HNE and αChy. Kinetic parameters were determined as described [20,21]. The active sites of HNE and αChy were titrated using a procedure based on the measurement of pNa released after enzymatic cleavage of MeOSucAlaAlaProValNA and SucAlaAlaProPheNA, respectively, at 37°C [22]. Product formation was monitored spectrophotometrically at a wavelength of 405 nm using a Bio Rad Microplate Reader model 3550. To titrate the different forms of α1-ATs (α1-AT, deglycosylated α1-AT and Z α1-AT), 7.5 nM HNE was incubated for 15 min at 37°C with 0–100 nM inhibitor, in the presence of 2 mM MeOSucAlaAlaProValNA. All following kinetic experiments were derived from α1-ATs and SP-A/α1-ATs complexes (obtained by incubation of α1-ATs, from 0 to 25 nM, with SP-A 15, 7.5, 1.5, 0.15 mM for 15 min at 37°C). See additional file 1for more details.

Results

To investigate the interaction between SP-A and α1-AT we studied whether Kass values, derived from incubating HNE with α1-AT, were modified by SP-A. Indeed we found a progressive decrease in the Kass as the SP-A concentrations increased (Table 1), irrespective of the animal source of SP-A. To exclude that the observed effect was due to LPS co-purified with SP-A [23], we repeated the assay using endotoxin-free SPA, but found no differences with native SP-A (Table 1). To reinforce this finding, in separate experiments we spiked α1-AT and SP-A/α1-AT mixtures with increasing amounts of LPS, without measurable effect on the Kass of α1-AT or SP-A/α1-AT mixture (not shown). As expected, [24], we found that the Kass of Z α1-AT for HNE was 3.5 fold lower than that of the normal, M α1-AT. When Z α1-AT was coupled with increasing SP-A concentrations, a further decrease in Kass towards HNE was observed (Table 2).

Table 1

Association rate constant (Kass M-1sec-1) for inhibition of HNE by α1-AT with SP-A. SP-A employed was both from humans affected by PAP or from rabbit, polymyxin treated and polymyxin-mannose treated. Data are means ± SD of three different experiments.

Reaction conditions		Kass, (M-1sec-1) means ± SD
		Human SP-A			Rabbit SP-A
α1-AT nM	SP-A nM	Native	Polymyxin-treated	Polymyxin/Mannose-treated	Native
7.5	0	3.40 ± 0.0079 × 107	3.40 ± 0.0079 × 107	3.40 ± 0.0079 × 107	3.40 ± 0.0079 × 107
7.5	0.15	1.84 ± 0.0577 × 107	1.84 ± 0.0580 × 107	1.86 ± 0.0565 × 107	1.82 ± 0.0585 × 107
7.5	1.5	1.70 ± 0.0623 × 107	1.70 ± 0.0631 × 107	1.72 ± 0.0618 × 107	1.68 ± 0.0620 × 107
7.5	7.5	5.20 ± 0.0483 × 106	5.22 ± 0.0480 × 106	5.00 ± 0.0478 × 106	5.40 ± 0.0490 × 106
7.5	15	4.30 ± 0.0513 × 106	4.30 ± 0.0520 × 106	4.30 ± 0.0520 × 106	4.40 ± 0.0498 × 106

Table 2

Association rate constant for inhibition of HNE by α1-AT and Z α1-AT with SP-A. Data are means ± SD of experiments performed in triplicate with 3 different batches of human SP-A and1batch of rabbit SP-A.

α1-AT/SPA ratio	α1-AT				Z α1-AT
	Reaction condition		Kass, (M-1sec-1)	decrease in Kass, n-fold	Reaction condition		Kass, (M-1sec-1)	decrease in Kass, n-fold
	α1-AT nM	SP-A nM			Z α1-AT nM	SP-A nM
	7.5	0	3.40 ± 0.0079 × 107	0	7.5	0	9.80 ± 0.0032 × 106	0
50	7.5	0.15	1.84 ± 0.0577 × 107	1.8	7.5	0.15	5.20 ± 0.0314 × 106	1.9
5	7.5	1.5	1.70 ± 0.0623 × 107	2	7.5	1.5	4.70 ± 0.0268 × 106	2.1
1	7.5	7.5	5.20 ± 0.0483 × 106	6.5	7.5	7.5	4.30 ± 0.0240 × 106	2.3
0.5	7.5	15	4.30 ± 0.0513 × 106	7.9	7.5	15	3.80 ± 0.0221 × 106	2.6

To exclude that the results were due to non-specific binding, we incubated 7.5 nM HNE with 0–100 nM α1-AT for 15 min at 37°C in microtiter plates or in glass tubes and then measured the residual HNE activity with 2 mM MeOSucAlaAlaProValNA, finding no difference between plastics and glass. Furthermore, to exclude binding of SP-A to plastics we incubated 15 nM SP-A with I125α1-AT (from 0 to 100 nM) at 37°C. The number of Cpm of the samples with SP-A were the same of wells without proteins. We concluded that our data were compatible with binding of α1-AT to SP-A.

Gel filtration HPLC was then used to determine the molecular weight of the SP-A/α1-AT complex. As shown in Figure 2A, profile a, a mixture of SP-A and α1-AT (1 mg/ml), gave two peaks, one corresponding to free 1-AT (unreacted α1-AT) and one, with a theoretical molecular weight of 1,642 kD (1-AT/SP-A complex), possibly corresponding to a complex made by one molecule of SP-A (670 kD) and 18 molecules of α1-AT (54 kD), suggesting that, under the experimental conditions applied, each monomer of SP-A bound one molecule of α1-AT. Further evidence that the first peak of profile a (Figure 2A) contained the complex SP-A/α1-AT was obtained by using an immunochemical assay in which a polypropylene plate was probed with antisera anti α1-AT and anti SP-A. As shown in Figure 2B, the first peak in profile a of Figure 2A contained both α1-AT and SP-A.

Figure 2

Isolation and immunodetection of the α. A: Isolation of the complex by gel filtration chromatography on two Biosep SEC – S 4000 columns connected in series using HPLC. Gel filtration profiles: commercial α1-AT (in profile c; 19.32 ± 0.1 mL); purified SP-A (in profile b; 16.49 ± 0.07 mL); α1-AT /SP-A complex (in profile a; 15.31 ± 0.04 mL) and unreacted α1-AT (in profile a; 19.35 ± 0.09 mL). Inset: calibration curve obtained using the following standards: A = α2-macroglobulin (725 kDa), B = aldolase (158 kDa), C = bovine serum albumin (67 kDa), D = chymotrypsinogen (25 kDa), E = cytocrome C (12.5 kDa). B: Immunodetection of the complex. α1-AT was added to wells a1 and a2, peak 1 (α1-AT /SP-A complex) of Figure 2A was added to wells b1 and b2, and SP-A to wells c1 and c2. Antiserum anti-α1-AT was added to wells a1, b1 and c1, antiserum anti-SP-A was added to wells a2, b2 and c2. Peak 1 (α1-AT /SP-A complex) is recognized by both antisera.

The effect of SP-A on the Kass of α1-AT for HNE was calcium-dependent, being abrogated by EDTA (Figure 3). Since the calcium-binding domain of SP-A lays at the COOH terminus, next to the CRD [25], we supposed that this part of SP-A could be involved in the binding of SP-A to α1-AT, via the α1-AT carbohydrate chains. Consistent with these findings, the addition of 1 M mannopyranoside to the SP-A/α1-AT mixture almost totally reversed the reduction in the Kass (Figure 3), most likely by interfering with the binding of CRD to α1-AT carbohydrate chains [26,27]. The fact that the lipid recognition domain of SP-A is located in the neck region of the molecule, far from the CRD [23], could explain the lack of influence of LPS on the binding of SP-A to α1-AT (Table 1).

Figure 3

Effects of calcium removal and sugar addition on K. Inhibition of HNE by α1-AT (7.5 nM), alone or coupled with 7.5 nM SP-A: (1) α1-AT alone, (2) α1-AT plus SP-A, (3) α1-AT plus SP-A with 5 mM EDTA and (4) α1-AT plus SP-A with 1 M MNOCH3. (Data are means ± SD of experiments performed in triplicate).

To better clarify the role of the CRD in the binding of SP-A to α1-AT, we modified both proteins by enzymatic digestion, deglycosylation or boiling and then used them to calculate the Kass of α1-AT for HNE and to deduce the molar parts of α1-AT bound to SP-A from the number of turnovers per inactivation of α1-AT not bound to SP-A. Thus we found that the CRD of SP-A appears to contain all the putative SP-A binding sites for α1-AT since, when incubated with α1-AT, it retained the same Kass, as that of native SP-A (Figure 4).

Figure 4

. Kass data are means ± SD of experiments performed in triplicate. SI values of the associations in bold at the top of the figure.

Turnover per inactivation (also referred to as stoichiometry of inhibition (SI) or partition ratio + 1) defines the number of moles of irreversible inhibitor required to completely inhibit 1 mole of target proteinase. The turnover number resulting from the interaction between unmodified SP-A and α1-AT was 24, i.e. one part of SP-A binds 23 molar parts of α1-AT and 24 SP-A plus α1-AT binds inhibit 1 part of enzyme (Figure 5). The same binding pattern emerged when Z α1-AT was used instead of α1-AT, suggesting that the difference in the Kass between the two variants of α1-AT is independent of the number of molar parts of inhibitor bound to SP-A.

Figure 5

Turnover numbers per inactivation. Turnover numbers were determined plotting residual enzyme activity/initial enzyme activity versus initial inhibitor concentration/initial enzyme activity. A: (1) native α1-AT, (2) deglycosylated α1-AT, (3) Z α1-AT, (4) native α1-AT coupled with deglycosylated and boiled SP-A, (5) deglycosylated α1-AT coupled with deglycosylated SP-A, (6) deglycosylated α1-AT coupled with deglycosylated and boiled SP-A, (7) native α1-AT coupled with boiled SP-A and (8)native α1-AT coupled with SP-A sugar chains. B: (9) native α1-AT coupled with native SP-A and (10) Z α1-AT coupled with native SP-A, (11) deglycosylated α1-AT coupled with native SP-A, (12) deglycosylated α1-AT coupled with boiled SP-A, (13) deglycosylated α1-AT coupled with SP-A sugar chains and (14) native α1-AT coupled with deglycosylated SP-A.

Deglycosylated α1-AT retains its ability to inhibit HNE (Kass 3.38 × 107 M-1sec-1). We did, however, find that the inhibitory activity of α1-AT is greatly decreased in the presence of SP-A (Kass 1.1 × 107 M-1sec-1, Figure 4), indicating that binding of SP-A to the carbohydrate moiety of α1-AT is not the only mechanism involved. The turnover number of the SP-A/deglycosylated α1-AT is 12, half that displayed by native α1-AT (Figure 4, 5). To explore other mechanisms of binding between SP-A and α1-AT, we incubated boiled SP-A and α1-AT. We found that boiled SP-A/native α1-AT displayed the same Kass and the same turnover number as native SP-A/deglycosylated α1-AT (Figures 4, 5). We postulated that SP-A carbohydrate chains could bind α1-AT, possibly through the amino acid backbone. In fact, carbohydrate chains isolated from SP-A mixed with deglycosylated α1-AT resulted in the same Kass and turnover number as those of native SP-A/deglycosylated α1-AT (Figures 4, 5). Besides these mechanisms of binding of SP-A to α1-AT, a third mechanism, i.e. a carbohydrate/carbohydrate interaction, probably exists since boiled SP-A and native α1-AT displayed a Kass of 1.9 × 107 M-1sec-1 and ~6 turnovers (Figure 4, 5).

Finally, we studied the binding of deglycosylated SP-A to α1-AT. The Kass of native α1-AT mixed with deglycosylated SP-A was 1.2 × 107 M-1sec-1 and the turnover number 18 (Figure 4, 5). Absence of SP-A/α1-AT binding, i.e. Kass 3.4 × 107 M-1sec-1, and a turnover number of 1, was achieved by two combinations: 1) SP-A deglycosylated and boiled with native α1-AT, and 2) both proteins deglycosylated (Figures 4, 5). In the former case, absence of SP-A carbohydrates and denaturation of CRDs hindered any possible binding of SP-A to native α1-AT. In the latter case, the binding was precluded by the absence of carbohydrates on both proteins, in spite of the presence of intact CRDs in the SP-A.

Discussion

The present data provide evidence for an in vitro interaction between SP-A and α1-AT. These glycoproteins belong to two systems of the lung that are supposed to act independently: the surfactant system and the proteinase/proteinase inhibitor system. Nevertheless, evidence for possible links between the two systems does exist. As an example, it has been shown that SP-A may be digested by elastolytic enzymes [28,29], and that inhalation of α1-AT in patients with cystic fibrosis may result in an increase of SP-A levels in bronchoalveolar lavage fluid (BALf) [30]. In addition, SP-D induces the production of matrix metalloproteinases by human alveolar macrophages [31], whereas the cysteine proteinase cathepsin H is involved in the first N-terminal processing step of SP-C [32]. The two systems may therefore interact in the lungs, both in physiologic and in pathologic pathways. The concentration of SP-A in the BALf of normal subjects is estimated to be ~277 nM [33]. Since approximately 1% of total SP-A is present in the free form [8,9], its concentration in BALf would be ~2.8 nM. Given that the concentration of α1-AT is ~5 μM [34], we reasoned that the two glycoproteins have a good chance of coming into contact during their life cycle.

Indeed our in vitro experiments indicate that the interaction between SP-A and α1-AT results in binding between them. This binding, which is calcium-dependent, appears to be complex since it could involve binding between the CRD of SP-A and carbohydrates on α1-AT, binding between SP-A carbohydrates and the protein backbone of α1-AT, and binding between the carbohydrate chains of both proteins.

Turnover per inactivation suggests that one part of SP-A binds 23 molar parts of α1-AT. Nevertheless, SP-A binds 11 molar parts of deglycosylated, fully active α1-AT (Figure 4, 5), thus suggesting a possible binding of SP-A carbohydrate chains to the amino acid backbone of α1-AT. Asn, to which carbohydrates of the native glycoprotein are linked [35], is a likely candidate. This hypothesis was confirmed by the results obtained with boiled SP-A and with isolated SP-A carbohydrate chains (Figure 4, 5). In support of this hypothesis, it has been reported that the binding of SP-A to influenza virus [36], herpes virus type 1 infected cells [37], and M. tuberculosis [38], involves N-linked carbohydrate chains on SP-A. Interestingly, there may be multiple binding sites on individual micro-organisms [12].

Our experiments also suggest a possible carbohydrate/carbohydrate interaction between SP-A and α1-AT. Such a type of linkage has been shown to operate in the calcium-mediated homotypic interaction between two Lewis (Lex) determinants (Galβ1→4[Fucα1→3]GlcNAc) involved in cell adhesion during murine embryogenesis [39]. Interestingly Lex-Lex interactions appear to be calcium-dependent [40], by involving van der Waal forces. The fact that ultra-weak interactions are involved explains why this aspect is often underestimated [39-41].

It is difficult to postulate whether the three proposed mechanisms of binding take place simultaneously between native proteins. It may be that the CRD plays the main role and that the other two mechanisms are less important or take place only as artificial mechanisms once the proteins have been manipulated.

The binding with SP-A results in a decrease in the inhibition of HNE by α1-AT. There are several known mechanisms that could explain the inactivation of α1-AT. Beside the physiologic irreversible suicide substrate mechanism by which α1-AT inhibits HNE [42], α1-AT may also be inactivated by oxidation of methionine residue(s) located at or near the active site [22,23]. Another mechanism of α1-AT inactivation is proteolytic degradation at or near the active site by a number of host and non-host, mostly microbial, proteinases [42]. Whether these mechanisms may act in vivo, thereby contributing to the imbalance between proteinases and inhibitors in the pathogenesis and progression of pulmonary emphysema, is still a debated issue.

With respect to the inhibitory activity of α1-AT, that of Z α1-AT is further impaired by this latter's enhanced tendency to undergo spontaneous polymerization [2]. This phenomenon, also known as loop-sheet polymerization, likely accounts for why Z α1-AT is less efficient at inhibiting HNE, and has been demonstrated to be present in vivo, since Z α1-AT polymers have been detected in the BALf of Z α1-AT subjects with emphysema [45]. We found that SP-A binds Z α1-AT and that the binding further reduces the Kass, which is already impaired with respect to that of α1-AT. Were this binding to happen in vivo, it would further decrease the antiproteinase activity of Z α1-AT.

The mechanism by which SP-A binding interferes with the α1-AT inhibitory mechanism is open to speculation. α1-AT inactivation taking place in vitro upon interaction between the two glycoproteins seems to occur because of the functional slowdown of α1-AT in the presence of SP-A, the turnover number shifting from 1 to 24. After an initial, non-covalent, Michaelis-like complex, the reaction between α1-AT and HNE progresses, through an acyl-enzyme intermediate resulting from peptide bond hydrolysis, to either a loop-inserted covalent complex (inhibitory pathway) or a cleaved serpin and free proteinase (non-inhibitory or substrate pathway) [42]. The number of turnovers for native α1-AT is 1 (Figure 5), indicating that the reaction inhibitor-HNE progresses towards the inhibitory pathway on the other side (Figure 6A). The number of turnovers after the incubation of native α1-AT or Z α1-AT with native SP-A is 24 (Figure 5), thus indicating that for α1-AT bound to SP-A the inhibitory pathway is precluded, and that the reaction inhibitor – HNE progresses mostly through the substrate pathway (Figure 6B).

Figure 6

Hypothetical mechanism of SP-A interference with α. A: interaction of α1-AT (I) with HNE (E). After an initial non-covalent Michaelis-like complex (EI), the interaction progresses through a tetrahedral intermediate (EI ♠), forming a covalent acyl-enzyme intermediate (EI ♥). The substrate pathway results in free HNE and cleaved α1-AT (I*); the inhibitory pathway results in a, about 100%, kinetically trapped loop-inserted covalent complex (E-I*). B: the SP-A (here shown as a trimer) interacts with α1-AT. In the presence of HNE, the formation of a covalent complex E-I* almost precluded (about 4%), and the reaction progresses through the substrate pathway towards free E and I* (cleaved α1-AT) – SP-A (96%). SI = stoichiometry of inhibition

In spite of the detailed dissection of the binding mechanism of SP-A to α1-AT in vitro, an obvious limitation of the present paper is the lack of specific studies investigating a possible interaction between SP-A and α1-AT in vivo. Nevertheless, some indirect evidence suggesting that such an interaction might take place is available, although it is not possible to address a plausible expectation of physiologic or pathophysiologic relevance of these findings. For example, a recent report has shown that in human sputum supramolecular complexes with heparan sulfate/Syndecan-1 and proteinase and inhibitors are present [46]. These complexes contain the proteinase inhibitors SLPI and α1-AT, NE as well, whose proteolytic activity is however not decreased . The large MW of SP-A makes difficult to highlight the occurrence of such supramolecular complexes including α1-AT by standard techniques [47]. Nevertheless, a report focusing on two-dimensional electrophoretic characteristics of BALf proteins in subjects affected by interstitial lung diseases [48] has intriguingly shown that some α1-AT fragments were superimposed on spots of SP-A, in its upper, acidic position. These findings, confirmed by mass spectrometric MALDI-TOF analysis, would suggest a possible SP-A/α1-AT interaction taking place in vivo.

Conclusion

We have shown that SP-A binds α1-AT, and that this binding results in a significant decrease in the association rate constant of α1-AT for HNE. The mechanism of the binding seems to be predominantly mediated by the SP-A CRDs, as indicated by the calcium dependence and by the turnovers for inactivation, but other mechanisms may be involved, such as an interaction between SP-A carbohydrates and the α1-AT amino acid backbone or between carbohydrate chains of both glycoproteins. The presence of these complex binding mechanisms would exclude the hypothesis that the α1-AT inhibition occurred simply due to steric inhibition of the large SP-A molecule, but it would rather suggest a programmed, coordinated mechanism.

The in vitro interaction described here, if present in vivo, would be a novel mechanism of impairment of α1-AT inhibitory activity. It might represent a physiologic mechanism of regulating α1-AT activity, especially in acute conditions (for example during defense against infections agents) [49], in which an excess of α1-AT would interfere with the physiologic role of proteinases. α1-AT is indeed a highly specialised proteinase inhibitor [50], but the presence in nature of several, robust mechanisms of α1-AT downregulation (i.e. inherited deficiency, susceptibility to oxidative stress and proteolysis, polymerization) would imply the occurrence of intrinsic risks related to the overexpression of a nearly perfect and immortal inhibitor. Therefore, the formation of supramolecular complexes SP-A/α1-AT might be a sort of reserve mechanism, taking place in case of need.

On the other hand, the interaction with SP-A would be of particular relevance in the pathogenesis of pulmonary emphysema associated with α1-AT deficiency, since it would contribute significantly to the complex mechanisms of imbalance between Z α1-AT and HNE in the lungs. Obviously, all these speculations need further investigations, first of all to understand whether or not SP-A/α1-AT binding is a relevant down-regulatory mechanism of α1-AT inhibitory activity in vivo.

Abbreviations

α1-AT, α1-antitrypsin

αChy, α chymotrypsin

(CH3)2SO, dimethylsulfoxide

CRD, carbohydrate recognition domain

CRF, collagenase-resistant fragment

HNE, human neutrophil elastase

MNOCH3, methyl-α-D-mannopyranoside

PNA, p-nitroanilide

SP-A, surfactant protein-A

Competing interests

The author(s) declare that they have no competing interests.

Authors' contributions

MG participated in the study design, performed most experiments, and helped to draft the manuscript. AL participated in the deglycosylation experiments and carbohydrate chains isolation. PI designed the experiments for the α1-AT/SP-A complex identification, and helped to draft the manuscript. CDS participated in the kinetic studies. PR performed the experiments for the α1-AT/SP-A complex identification. DD performed the purification of SP-A and CRF. NC took part to some kinetic experiments. EP participated in the coordination of the study. AB performed the purification of SP-A and CRF, helped to draft the manuscript and critically reviewed it. ML conceived the study, participated in its design, and coordinated the manuscript final version. All authors read and approved the final manuscript.

Additional File 1

contain portion of Methods' section and include details on modification of native proteins used in the experiments and details of kinetic procedures.

Click here for file