Kinetics of nonproteolytic incorporation of a protein ligand into thermally activated alpha 2-macroglobulin: evidence for a novel nascent state.

We have previously shown that antigens complexed to the receptor-recognized form of alpha(2)-macroglobulin (alpha(2)M*) demonstrate enhanced immune responsiveness mediated by the low density lipoprotein receptor-related protein LRP/CD91. Recently, we developed a proteinase-independent method to covalently bind antigens to alpha(2)M*. Given the potential applications of this chemistry, we analyzed the kinetics, thermodynamics, and pH dependence of this reaction. The incorporation of lysozyme into alpha(2)M* was a mixed bimolecular second-order reaction with a specific rate constant of 91.0 +/- 6.9 m(-1) s(-1), 50.0 degrees C, pH 7.4. The activation energy, activation entropy, and Gibbs' free energy at 50.0 degrees C were 156 kJ mol(-1), 266 J mol(-1) K(-1), and 70 kJ mol(-1), respectively. The rate of incorporation increased as a function of pH from pH 5.0 to 7.0 and was unchanged thereafter. Furthermore, the reaction between alpha(2)M* and lysozyme was irreversible. The data are consistent with a two-step mechanism. In the first step, alpha(2)M* reforms its thiol ester bond, entering a reactive state that mimics the proteolytically induced "nascent state." In the rate-limiting second step, the reformed bond quickly undergoes nucleophilic attack by lysozyme. The kinetic equations derived in this study are the basis for optimizing the formation of stable alpha(2)M*.antigen complexes.

Human ␣ 2 -macroglobulin (␣ 2 M) 1 is a 720-kDa tetrameric glycoprotein that can inhibit proteinases of all classes and specificities through a unique mechanism involving both steric entrapment and covalent binding (1). ␣ 2 M is composed of four identical subunits that form a cage-like structure (2). Each subunit contains a "bait region," a stretch of amino acids that is highly susceptible to proteinase cleavage (3), and a functionally important internal ␤-cysteinyl-␥-glutamyl thiol ester (4 -6). In native ␣ 2 M, the thiol ester bond is composed of the side chains of Cys 949 and Glx 952 and appears to sit in a protective hydrophobic pocket that restricts access of most nucleophiles (4 -6). Limited proteolysis of the bait region activates the thiol ester, resulting in a transient "nascent state" in which the thiol ester bond is extremely susceptible to nucleophilic substitution. Sub-sequently, a major conformational change occurs in ␣ 2 M, yielding the more compact receptor-recognized form (␣ 2 M*) (7). The covalent attachment of the proteinase to ␣ 2 M occurs between Glx 952 of the nascent state thiol ester and an available lysine or NH 2 terminus of the proteinase. Other proteins, if present during proteolysis, can react with the thiol ester and may compete with the proteinase for covalent binding (8,9).
Although traditionally viewed as a proteinase inhibitor, ␣ 2 M is also a major serum carrier of cytokines and growth factors and has been implicated in the regulation of the immune system. Our laboratory has demonstrated that covalently linking an antigen to ␣ 2 M* enhances uptake, processing, and presentation of that antigen by peritoneal macrophages (10). Macrophages pulsed with lysozyme complexed to ␣ 2 M* achieved effective presentation to T cells with 200-to 250-fold less protein than cells pulsed with free lysozyme. This enhancement was mediated by the ␣ 2 M* receptor, the low density lipoprotein receptor-related protein LRP/CD91. These studies were extended in vivo by demonstrating that, compared with free lysozyme, lysozyme complexed to ␣ 2 M* resulted in 10-to 500-fold higher IgG titers (11). These levels were similar to those elicited by emulsification in Complete Freund's Adjuvant, generally considered the most potent adjuvant currently available.
The ␣ 2 M*⅐antigen complexes were generated by co-incubating an antigen with ␣ 2 M* in the presence of a proteinase (12). This method for complex formation results in complexes that contain varying amounts of proteinase that compete with the antigen for binding to the thiol ester. Although some antigen is covalently incorporated inside ␣ 2 M*, the trapped proteinase remains functional and is sometimes able to destroy the antigen. 2 Additionally, if these complexes are used for induction of an immune response, they may lead to the generation of antibodies against the proteinase (11).
We have recently developed a method for incorporating nonproteolytic proteins into ␣ 2 M* without the use of a proteinase (13). We observed that NH 3 or CH 3 NH 2 -activated ␣ 2 M, when heated, can covalently bind to co-incubated proteins. The absence of the proteinase within the ␣ 2 M* cage allows more and larger antigens to incorporate into ␣ 2 M* (13). 3 More recent studies from our laboratory using complexes generated in this manner have demonstrated a remarkable increase in both the humoral and T-cell-mediated immunogenicity against the hepatitis B surface antigen and the HIV peptide C4-V3 (14). 4 As a result of the recent interest in LRP/CD91-mediated antigen presentation, the adjuvant-like properties of ␣ 2 M* have been receiving increasing attention. The optimization of ␣ 2 M*⅐antigen complex formation hinges upon the chemistry of this reaction. Recent attempts to create these complexes using nonproteolytic activation, however, have resulted in poor antigen incorporation (15). Detailed kinetic analysis presented here allows for the optimization of antigen incorporation into ␣ 2 M*. By appropriate use of this chemistry, significant amounts of antigen can be incorporated into ␣ 2 M*.

EXPERIMENTAL PROCEDURES
Preparation of ␣ 2 M and 125 I-HEL-Human ␣ 2 M was purified from frozen human plasma (American Red Cross, Charlotte, NC) according to a previously published protocol (16). The concentration of ␣ 2 M was determined using A 280 (1%/1 cm) ϭ 8.93 and a molecular mass of 720 kDa (17). Two commercial ␣ 2 M preparations were purchased from Sigma-Aldrich. The thiol ester-cleaved ␣ 2 M* derivative was prepared by incubating native ␣ 2 M with 0.2 M NH 4 CO 3 , pH 8.0, for 60 min at room temperature. Excess NH 3 was removed by size-exclusion chromatography on Sephadex G-25 (Amersham Pharmacia Biotech) equilibrated in 0.1 M NaPi and 0.15 M NaCl (PBS), pH 7.4. The efficiency of the conversion to ␣ 2 M* was determined by following its electrophoretic mobility on nondenaturing, nonreducing polyacrylamide 4 -15% Tris-HCl gels (Bio-Rad, Hercules, CA) (7). HEL (Sigma-Aldrich) was iodinated with Iodogen-coated tubes (Pierce) and Na 125 I (PerkinElmer Life Sciences). Typically, 2.5 mg of HEL was incubated with 1.0 mCi of 125 I in 500 l of PBS, pH 7.4, for 30 min at room temperature. Residual 125 I was removed on Sephadex G-25 equilibrated in PBS, pH 7.4. The concentration of lysozyme was determined using A 280 (1%/1 cm) ϭ 26.5, assuming a molecular mass of 14 kDa (18). Both ␣ 2 M* and 125 I-HEL were sterile-filtered through Millex-GV 0.22 m filter units (Millipore) and stored at 4°C in PBS, pH 7.4. Proteins were used within 2 weeks of conversion (␣ 2 M*) or iodination ( 125 I-HEL).
Measurement of 125 I-HEL Incorporation into ␣ 2 M*-The kinetics of ligand incorporation were measured using 125 I-HEL as a model ligand, as it has been in previous studies involving the incorporation of ligands into ␣ 2 M with and without the use of proteinase (10,11,13). All measurements of kinetics were performed during the first 10 min of the reaction, whereas experiments requiring reaction completion were performed for 180 min. Varying amounts of ␣ 2 M* (0.025-0.2 M) were incubated with 125 I-HEL (2.5-20 M) at 50.0°C for the appropriate time (2.5-180 min) in PBS, pH 7.4. Incubations were performed in a water bath, the temperature of which was monitored with a digital thermometer (Control Company, Friendswood, TX) with an accuracy of Ϯ0.1°C. After incubation, reaction mixtures were subjected to nondenaturing, nonreducing, pore-limit polyacrylamide gel electrophoresis and nonreducing SDS-polyacrylamide gel electrophoresis to determine the amount of total and covalent 125 I-HEL that was bound to ␣ 2 M*, respectively. Electrophoresis was performed on 4 -15% Tris-HCl gels (Bio-Rad) that were stained, dried, and analyzed on a Molecular Dynamics STORM 860 (Sunnyvale, CA). 125 I-HEL incorporation into ␣ 2 M* was analyzed by measuring the amount of 125 I-HEL associated with the band corresponding to the molecular weight of ␣ 2 M* using ImageQuant 5.1 (Molecular Dynamics). Linear and nonlinear regressions were performed using Jandel-Scientific SigmaPlot (Chicago, IL). All noncovalent 125 I-HEL binding (determined by subtracting the amount of 125 I-HEL covalently bound from the total 125 I-HEL bound) was determined to be nonspecific due to its linear, nonsaturable nature.
To determine whether the approach employed to terminate the reaction affected the binding kinetics, two different methods were compared. Reactions were stopped either by cooling on ice or by adding SDS at ϳ100°C and immediately incubating the mixture at 100°C for 10 min. No difference in the rate of 125 I-HEL incorporation into ␣ 2 M* was observed, and all subsequent reactions were stopped on ice.
Temperature and pH Studies of 125 I-HEL Incorporation into ␣ 2 M*-The temperature dependence of the rate of incorporation of 125 I-HEL into ␣ 2 M* was determined between 47.5°C and 55.0°C at pH 7.
where k is the specific rate constant, A is the pre-exponential factor, and R is the gas constant (8.314 J/mol K). The activation entropy (S) was determined using the following equation: where k B is the Boltzman constant (1.38 ϫ 10 Ϫ23 J/K), h is Planck's constant (6.626 ϫ 10 Ϫ34 J/s), and T m is the average temperature. The Gibbs' free energy (⌬G ‡ ) at 50.0°C was analyzed with the equation below (⌬H ‡ , activation energy; ⌬S ‡ , activation entropy).
The effect of pH on the reaction rate of 125 I-HEL incorporation into ␣ 2 M* was studied between pH 5 and pH 9.

Time and Concentration
Dependence of 125 I-HEL Incorporation into ␣ 2 M*-Studies described below indicated that the incorporation of 125 I-HEL into ␣ 2 M* at 180 min saturated at a 125 I-HEL:␣ 2 M* ratio of ϳ100:1; therefore, binding was performed under these conditions to observe the reaction time course. The incorporation data fit a monophasic exponential curve that was linear for 15-20 min and approached equilibrium at 180 min (data not shown). The relationship between 125 I-HEL incorporation into ␣ 2 M* and the ratio of the concentrations of the reactants was then measured after 180 min of incubation at 50.0°C, with 125 I-HEL:␣ 2 M* molar ratios ranging from 1. ␣ 2 M* concentration (Fig. 2B), and the slope of the linear regression was 1.02 Ϯ 0.09, suggesting that the reaction was first order with respect to ␣ 2 M*. This procedure was repeated to determine the order of 125 I-HEL and yielded similar results (Fig. 2, C and D). The concentration of ␣ 2 M* was held constant at 0.2 M, whereas 125 I-HEL concentration varied from 2.5-20 M. The relationship between the logarithm of the reaction rate and 125 I-HEL concentration was also linear with a slope of 1.08 Ϯ 0.13, indicating that the reaction was first order with respect to 125 I-HEL. Using the rate equation and the orders of 125 I-HEL and ␣ 2 M*, the specific rate constant (k) at 50.0°C was calculated as 91.0 Ϯ 6.9 M Ϫ1 s Ϫ1 and was independent of the reactants' concentrations. Therefore, the experimentally derived rate equation for covalent incorporation of 125 I-HEL into ␣ 2 M* can be expressed by the relationship: in which ␣ 2 M* s represents an individual subunit of the ␣ 2 M tetramer.
Temperature and pH Dependence of 125 I-HEL Incorporation into ␣ 2 M*-The temperature dependence of the rate of 125 I-HEL incorporation into ␣ 2 M* was determined from 47.5°C to 55.0°C at pH 7.4. Samples were collected and analyzed at 2.5, 5.0, 7.5, and 10.0 min. Specific rate constants were computed at each temperature using the experimentally derived rate equation. The Arrhenius plot of the data demonstrates a linear relationship between ln k and the inverse of temperature within the limited temperature range studied ( Fig. 3 and Table  I). From the plot, the following thermodynamic parameters were calculated: ⌬H ‡ ϭ 156 kJ mol Ϫ1 , ⌬S ‡ ϭ 266 J mol Ϫ1 K Ϫ1 , and ⌬G ‡ at 50.0°C ϭ 70 kJ mol Ϫ1 .
Analysis of the influence of pH on the reaction rate of 125 I-HEL incorporation into ␣ 2 M* was studied in the pH range of 5.0 -9.0. The plots of the amount of 125 I-HEL incorporated versus time are linear at all pH values investigated during the time interval studied. The reaction rate increased as pH was raised from 5.0 to 7.0. No change was observed as pH increased from 7.0 to 9.0 (Table II) Stability of 125 I-HEL⅐␣ 2 M* Complexes-It has been observed that the reaction between the thiol ester and the nucleophiles ammonia and methylamine can be reversed at high temperatures (19). To determine whether nonproteolytic incorporation  a The data are a summary of the results presented in Fig. 3. was also reversible, 125 I-HEL⅐␣ 2 M* complexes were generated by incubating a 100-fold excess of 125 I-HEL with ␣ 2 M* for 180 min at 50.0°C. Free 125 I-HEL was removed by five washes through YM-100 filter units, and the purified complexes were heated to 50.0°C. Samples were collected at various times between 5 and 180 min, and the amount of 125 I-HEL complexed to ␣ 2 M* was compared with the total amount of 125 I-HEL complexed at time 0 (Fig. 4). After 180 min of incubation at 50.0°C, the amount of 125 I-HEL bound to ␣ 2 M* was unchanged, indicating that the ␣ 2 M*⅐ 125 I-HEL complexes were intact and that the incorporation was irreversible. This result is in marked contrast to the reversibility of the reaction of small nucleophiles incorporated into the thiol ester of ␣ 2 M* (19). DISCUSSION Our laboratory has previously demonstrated that ␣ 2 M* significantly enhances antigen processing and presentation by carrying antigen into macrophages through a receptor-mediated process involving LRP/CD91 (10). Compared with free HEL, there was a 5-fold increase in complex uptake but a 100 -1000-fold increase in T-cell response (10). These studies demonstrate that the enhanced responsiveness results from enhanced antigen presentation by antigen-presenting cells and is not simply dependent on an increase in ␣ 2 M*⅐HEL complex uptake. A potent B-cell response was observed in vivo with 10to 500-fold higher IgG titers toward HEL when it was complexed to ␣ 2 M* (11). We have recently developed a novel method of covalent attachment of ligands to ␣ 2 M* without the addition of proteinases. Studies have demonstrated that complexing antigens to ␣ 2 M* in this manner can increase the antigen-specific B-and T-cell responses. More recent studies by other investigators propose a broader role for LRP/CD91 in antigen presentation, suggesting that the receptor may also uptake and process antigens coupled to various heat shock proteins (20). When the available literature is considered, it is likely that ␣ 2 M* and other molecules carrying antigens employ LRP/CD91 as an entry point for antigen delivery through multiple pathways.

FIG. 2. Determination of reaction order with respect to 125 I-HEL and
In general, host response toward protein subunit antigens is poor. It is therefore likely that ␣ 2 M* will be employed to produce complexes with candidate subunit vaccine antigens to boost the responsiveness of human recipients toward these preparations. The ␣ 2 M* employed for such purposes will not be generated by proteinases for a variety of reasons. First, to incorporate antigens into ␣ 2 M*, the antigen must be present during the activation step. Nonproteolytic proteins can only form covalent adducts if they are present during the nascent state resultant from attack of the proteinase on the bait region of ␣ 2 M. (12) Second, when proteinases are employed to activate and incorporate antigens into ␣ 2 M, the proteinase within ␣ 2 M* is active and can cleave co-incorporated proteins as noted above. Third, the presence of a co-incorporated proteinase within the ␣ 2 M*⅐antigen complex competes with the nonproteolytic protein for space and binding sites within ␣ 2 M*. This decreases the amount of antigen that can be incorporated (12). For all these reasons, the ␣ 2 M* used to prepare ␣ 2 M*⅐antigen complexes should ideally be prepared by the reaction of ␣ 2 M with NH 3 . This preparation suffers from none of the disadvantages cited above. Moreover, this preparation can be made in advance and stored for long periods of time until required to produce an ␣ 2 M*⅐antigen complex. Given the increased interest in employing ␣ 2 M* to target antigens for presentation, it is important to understand the chemistry of the reaction between nonproteolytic proteins and ␣ 2 M-NH 2 at elevated temperatures. To date, this reaction has not been characterized in any detail.
The current studies were undertaken to address these issues. It is demonstrated that the reaction rate is equally and proportionally dependent upon the concentrations of both the ligand and ␣ 2 M*, indicating that the incorporation reaction is a mixed, bimolecular reaction occurring under second-order conditions. The absence of both a lag phase and a biphasic reaction suggests a lack of cooperativity among the four thiol esters. The finding that this reaction is bimolecular is consistent with the binding of HEL occurring through nucleophilic substitution. This is supported by the finding that nucleophilic attack is the mechanism by which proteinases and small primary amines bind to ␣ 2 M (4, 5, 21). The specific rate constant for HEL incorporation into ␣ 2 M* is significantly higher than the specific rate constants of nucleophilic attack by uncharged amines on native ␣ 2 M. Incorporation of HEL is ϳ64 times faster than thiol ester cleavage by ammonia and Ͼ1000 times faster than cleavage with isopropylamine (21). This may result from the difference in thiol ester exposure or availability in the nascent state between these reactions. Nonproteolytic incorporation occurs by way of an exposed, reformed thiol ester; by contrast, primary amines are sterically hindered from reacting with the thiol ester in native ␣ 2 M, given its location in a hydrophobic pocket (4 -6). Additionally, the large activation entropy of the reaction may contribute to the magnitude of the specific rate constant due to its effect on the free energy of the reaction.
The data for both pH and temperature also support nucleophilic substitution as the reaction mechanism. The increase in reaction rate as pH rises to 7.0 is consistent with an increase in reactivity of attacking nucleophiles, such as amines and sulfhydryls, as a consequence of deprotonation. Additionally, the  consistency of the pH data and the activation energy for ligand incorporation with data generated on spontaneous thiol ester bond cleavage using hexapeptide models of thiol esters suggests that the thiol ester has reformed and is available for nucleophilic attack (22).
The large activation energy is consistent with our initial studies on the binding of nonproteolytic ligands to ␣ 2 M* (13) and reflects the thermodynamic unfavorability of thiol ester reformation. It has been observed that when ␣ 2 M* is heated in the absence of a ligand, the thiol esters regenerate, and the molecule reverts to the native form (19). We hypothesize that heated ␣ 2 M* passes through a state that mimics the proteolytically induced nascent state, such that nonproteolytic ligands can be incorporated. This is supported by the fact that during the conversion to ␣ 2 M*, native ␣ 2 M can only bind ligands while in the nascent state, in which the labile thiol esters are briefly exposed to ligands (4,5).
Based on this data, we propose that the incorporation of ligands into ␣ 2 M* occurs through a two-step reaction (Fig. 5) shown below.
Our model incorporates a nascent-like state, Ͻ␣ 2 MϾ, that appears to be the transition state necessary for ligand binding. We propose that the thiol ester reforms but is then quickly subjected to nucleophilic attack by the ligand. Binding of the thiol ester, in the nascent-like state, to the ligand drives the formation of the ␣ 2 M*⅐ligand complex instead of allowing the nascent-like state to proceed back to native ␣ 2 M. The nascent state is considered to be a transient intermediate, and the inability to detect it when the reaction mixture is separated electrophoretically attests to its high reactivity. We suggest that only the first step of the reaction requires the addition of heat; in the nascent state, energy is not required for ligand binding, as shown by the ability of native ␣ 2 M to react with and bind proteinases at 4°C (1). This model addresses the reversibility of nascent-state formation shown by Gron et al. (19) and the irreversibility of ligand incorporation. This irreversibility may be due to stabilization of the ligand within the ␣ 2 M* cage-like structure through protein-protein interactions. This proposed mechanism is consistent with the experimentally derived rate equation, with Step II as the rate-limiting reaction and the specific rate constant (k) equal to (k 1 k 2 )/k Ϫ1 . The Gron and Pizzo method of ␣ 2 M*⅐antigen complex generation has been used by others to suggest that uptake by LRP/ CD91 can result in antigen presentation on major histocompatibility complex class I (15). However, caution is required when interpreting this previous study. The ␣ 2 M employed was obtained commercially, and the incubation of ␣ 2 M with the antigen was performed for only 10 min at 50°C. The form of ␣ 2 M in the commercial preparation was not stated. It is our experience that commercial preparations show varying ratios of ␣ 2 M: ␣ 2 M*. In the extreme case, preparations consist predominately of one form or the other. Because native ␣ 2 M cannot incorporate antigens by heating, the ratio of ␣ 2 M:␣ 2 M* in such preparations is critical. Moreover, proteolytically activated forms of ␣ 2 M* do not incorporate antigen. Thus, unless a commercial preparation of ␣ 2 M:␣ 2 M* contains ␣ 2 M-NH 2 , it will not be suitable for heat-induced generation of ␣ 2 M*⅐antigen complexes. Assuming that the commercial preparations employed in these studies were predominately ␣ 2 M*-NH 2 , the conditions under which ␣ 2 M*⅐antigen complexes were produced are also critical. In the current study, we have derived equations that can be employed to calculate the amount of ␣ 2 M*⅐antigen formed for any given time of incubation. The 10-min incubation FIG. 5. Model of ligand incorporation into ␣ 2 M*. ␣ 2 M*-NH 3 (␣ 2 M* derived with ammonia) converts to Ͻ␣ 2 M*Ͼ with the addition of heat in a reversible manner. In the presence of excess ligand, Ͻ␣ 2 M*Ͼ binds to the ligand and forms a covalent complex. Small amounts of Ͻ␣ 2 M*Ͼ may convert to the native form if Ͻ␣ 2 M*Ͼ is not subjected to nucleophilic attack by ligand while the thiol esters are exposed. period employed in the above cited study is predicted to yield 0.8% ␣ 2 M*⅐antigen complex; however, only 0.1% complex production was reported (15). Optimal reaction conditions, as defined in the present study, will make it possible to obtain complex preparations in which every ␣ 2 M* molecule carries multiple antigens. Such complexes can be expected to have significant effects on antigen presentation mediated by ␣ 2 M*.