Gas Phase Characterization of the Noncovalent Quaternary Structure of Cholera Toxin and the Cholera Toxin B Subunit Pentamer

Abstract

Cholera toxin (CTx) is an AB₅ cytotonic protein that has medical relevance in cholera and as a novel mucosal adjuvant. Here, we report an analysis of the noncovalent homopentameric complex of CTx B chain (CTx B₅) using electrospray ionization triple quadrupole mass spectrometry and tandem mass spectrometry and the analysis of the noncovalent hexameric holotoxin usingelectrospray ionization time-of-flight mass spectrometry over a range of pH values that correlate with those encountered by this toxin after cellular uptake. We show that noncovalent interactions within the toxin assemblies were maintained under both acidic and neutral conditions in the gas phase. However, unlike the related Escherichia coli Shiga-like toxin B₅ pentamer (SLTx B), the CTx B₅ pentamer was stable at low pH, indicating that additional interactions must be present within the latter. Structural comparison of the CTx B monomer interface reveals an additional α-helix that is absent in the SLTx B monomer. In silico energy calculations support interactions between this helix and the adjacent monomer. These data provide insight into the apparent stabilization of CTx B relative to SLTx B.

INTRODUCTION

Cholera toxin (CTx), produced by toxinogenic strains of Vibrio cholerae, is a heterohexameric AB₅ protein belonging to the family of related toxins that are associated with human disease. Infection with V. cholerae can lead to an acute gastrointestinal disease resulting in toxemia and chronic diarrhea causing the loss of over 10–20 liters of body fluids per day in some cases (1) and frequently resulting in a mortality of 20–50% in untreated cases (2,3). Indeed, there are over 200,000 reported new cases of cholera each year (4). CTx, the main toxic product produced during infection, has been characterized, along with the producing bacterium, as a potential ‘biowarfare’ agent by the US and UK defense agencies (5). The structure of CTx holotoxin and the CTx B pentameric subunits have been determined by means of x-ray crystallography (6–8). These studies revealed that the five B subunits (∼11.6 kDa/subunit) form a doughnut-shaped ring and that the A subunit (∼27.5 kDa) is composed of two distinct domains, A1 and A2. The A1-domain is responsible for the enzymatic activity of the toxin, whereas the A2-domain assists in the tethering of the A1-domian to the B subunit pentamer by inserting noncovalently into and through the central pore of the B₅ complex (Fig. 1). The CTx A1 subunit is an ADP-ribosyltransferase and NAD-glycohydrolase that can modify G proteins by ADP ribosylation (9). The main intracellular target of CTx directly associated with the induction of fluid secretion is Gs, a G protein involved in the activation of the adenylate cyclase complex (10–12), although the toxin can affect numerous metabolic processes through the modification of other G proteins. Modification of Gs by CTx leads to constitutive activation of the adenylate cyclase complex and elevated intracellular cAMP levels (13–15), which in turn stimulates active chloride secretion (16–18) resulting in a severe form of secretory diarrhea.

FIGURE 1.

Crystal structure of CTx and the CTx B₅ pentamer. (A) CTx holotoxin viewed sideways with the surface of the B pentamer that contains the GM1 binding sites facing downward. The catalytic A chain showing secondary structure. The A2 fragment can be seen to insert into the core of the B pentamer. (B) CTx B₅ pentamer shown from the top, with a single monomer indicated.

To accomplish these intracellular changes, CTx must reach the cytosol of its target cells. CTx binds, via its B pentamer, to the glycosphingolipid GM1 ganglioside (19–21), which is present in the plasma membrane of virtually all cell types, including those cells that line the intestinal tract and certain cells associated with the control and function of the immune system (22). Binding of CTx to GM1 at the plasma membrane triggers toxin uptake and subsequent retrograde delivery of the toxin to the endoplasmic reticulum (ER) (23–28) where the A1-domain is reduced from the A2-B₅ complex and then retrotranslocated across the ER membrane into the target cell cytosol where it can modify its targets (27,29–32).

In addition to the well-characterized cytotonic properties of CTx, interaction between the CTx holotoxin or the isolated B₅ pentamer and specific target cells can also promote signaling pathways not linked to cAMP induction or indeed to other second messenger molecules. This can lead to different immunomodulatory functions, such as mucosal adjuvant effects, depending on the target cell type (22). A key element in the diverse signaling activities of this toxin is the initial oligomerization of GM1 receptors upon binding the B-chain pentamer. This is believed to promote the recruitment of both signaling and accessory molecules into cholesterol-enriched membrane microdomains (lipid rafts) to establish and enhance signaling cascades (22,33). As such, analysis of the noncovalent interactions that promote and maintain the quaternary structure of this toxin is of fundamental importance since such interactions have significant biochemical and medical implications.

There have been relatively few reports of the analysis of CTx by mass spectrometry (MS) (34,35). An in-depth study by liquid chromatography MS (LC-MS) and LC tandem MS (LC-MS/MS) provided sequence information derived from tryptic digestion and showed that the toxin was well suited to mass spectrometric detection (35). During this previous investigation, the AB₅ holotoxin was not detected. Here, we demonstrate that the complete noncovalent assembled AB₅ holotoxin remains intact during the electrospray ionization (ESI) process and can be observed in the mass spectrum.

ESI-MS has been shown to be an extremely useful tool in the study of the AB₅ noncovalent assemblies of the Shiga-like toxin (SLTx) and multivalent complexes of SLTx with globotriaoside (36). We have previously shown that SLTx from Escherichia coli can be detected over a pH range of 3–7 (37), in contrast to the previous study where the AB₅ holotoxin was not detected at neutral pH (36). Investigation of the dissociation of the homopentamer B₅, B₄, and B₃ assemblies by collision-induced dissociation (CID) in an instrument capable of performing MS/MS experiments was shown to provide dissociation pathways of the assemblies. Information relating to the structure of the assemblies was inferred from the product ions formed from the MS/MS experiments. The results obtained for the dissociation of the homopentamer B₅ during the previous study (37) were consistent with those obtained by thermal decomposition by blackbody infrared radiative dissociation within the cell of a Fourier transform ion cyclotron resonance mass spectrometer (38).

Here, ESI-MS and ESI-MS/MS have been successfully used to examine the noncovalent interactions that exist between the subunits of the complete CTx holotoxin and the pentameric B chains. The first objective of this investigation was to examine the ions produced from the complete hexameric noncovalent holotoxin complex (CTx AB₅). Our results clearly show that the intact complex was observed in the gas phase in a time-of-flight (TOF) mass spectrometer. Furthermore, we report the dissociation of the noncovalent homopentameric binding subunit of the CTx by means of CID in a collision cell of a triple quadrupole mass spectrometer. This has allowed us to probe the noncovalent assembly of CTx and CTx-B5s through their gas phase dissociation pathways.

MATERIALS AND METHODS

CTx and CTx B-subunit were purchased from Sigma (St. Louis, MO). The toxin samples were extensively dialyzed against 10 mM ammonium acetate (Sigma) at pH 7.0, desalted using PD-10 gel filtration columns, and concentrated to a final protein concentration of 16.3 μM for CTx holotoxin and 8.97 μM for CTx B in a Centricon 10 apparatus (Millipore, Watford, UK).

Mass spectrometry

Experiments were performed in a single TOF mass spectrometer (LCT, Waters MS Technologies, Manchester, UK) and a triple quadrupole mass spectrometer (Quattro Ultima, Manchester, UK). In each instrument, noncovalent assemblies were observed by increasing the backing pressure between the source and high vacuum region. Both instruments were equipped with the standard Z-spray ESI source and operated at a source and desolvation temperature of 110°C. Sample solutions were introduced into the source region of the instruments by conventional electrospray, operated in positive mode of ionization with a capillary voltage of 3 kV, and optimized for the transmission of noncovalent assemblies. In the triple quadrupole instrument, tandem mass spectrometry (MS/MS or MS²) was carried out using argon collision gas at a pressure of 2.5 × 10⁻³ mbar within the radio-frequency (RF)-only hexapole collision cell.

Data acquisition and processing were carried out using MassLynx (V3.5). Transform or, alternatively, maximum entropy-based software was used to find the masses of the subunits and hence the charges on each multiply charged ion. All spectra shown were subjected to minimal smoothing. The upper m/z range of the triple quadrupole instrument is 4000 so that, since the monomer has a mass of nearly 11,600, ions of the type could be observed only if y is 3x or greater.

The complex formed by the homopentameric binding subunit of the CTx (CTx B₅) and the hexameric holotoxin (CTx AB₅) was studied by direct infusion at a concentration of 4.5 pmol/μL and 8 pmol/μL in 10 mM aqueous ammonium acetate at a pH of 4.0–7.0, adjusted by addition of dilute ammonia or formic acid solution, respectively. A Harvard Apparatus (South Natick, MA) Model 22 syringe pump was used to provide a flow rate of 4 μL/min.

Molecular modeling

The structures of SLTx (PDB code 1BOV) and CTx B (PDB code 1FGB) were taken from the Protein Data Bank (PDB) at Brookhaven (39) and imported into the Molsoft ICM program (40). The coordinates of the structures were then converted into internal coordinate space to allow regularization of the structures to be performed. Interchain energy calculations were made using the standard ICM force field (40) between two adjacent monomers (chains A and B in 1BOV and chains D and E in 1FGB). Interactions were specifically analyzed and images made using ICM.

RESULTS AND DISCUSSION

Triple quadrupole MS for the AB₅ holotoxin

Initially, the CTx holotoxin (4 pmol/mL) was analyzed denatured in a solution of acetonitrile/H₂O + 0.2% formic acid. This confirmed the homogeneity of the CTx B components. However, the CTx A protein was heterogeneous and consisted of two components (Fig. 2). Maximum entropy provided the average molecular mass of 11605.2 Da for the CTx B subunit and average molecular masses of 26581.4 Da and 26905.8 Da for the heterogeneous CTx A protein. The maximum entropy deconvoluted spectrum obtained clearly identifies two different major components for the A protein (Fig. 2 B).

FIGURE 2.

ESI-triple quadrupole mass spectrum of the denatured cholera holotoxin at a concentration of 4 pmol/μL in acetonitrile/H₂O + 0.2% formic acid (A). Maximum entropy deconvoluted spectrum expanded in the region from m/z 10,000 to 30,000 highlighting the heterogeneity of the A unit (B).

The A-subunit consists of two chains, A1 (194 residues) and A2 (46 residues) which are connected by a disulphide bridge, whereas the B chain (103 residues) is a single polypeptide (Fig. 3 A). The heterogeneity observed within the A chains of these family of proteins is not uncommon, as previously reported for the related AB₅ protein, SLTx (37). The average mass observed at 26905.8 Da for CTx A chain is identical to one observed previously (35) and has been ascribed to the loss of methionine from the N-terminal of the A2 chain together with the loss of two serine residues from the C-terminus of the A1 chain (Fig. 3 B). The average mass difference between the observed and predicted mass of 26904.8 Da is 41 parts-per-million (ppm). The peak at 26581.4 Da was not previously observed and probably corresponds to the loss of methionine from the N-terminus of the A2 chain together with the loss of two serine residues, arginine, proline, and alanine from the C-terminus of the A1 chain (Fig. 3 C). Again the average mass difference between the observed and predicted mass of 26580.3 Da is 41 ppm. The difference between the predicted and observed mass was extremely low in both cases, conferring a high degree of confidence in the predicted cleavage patterns. Although the masses agree well with theory, this does not necessarily negate the need to determine the exact sequences for samples in which the origin is not precisely known. However, in this case, the samples were obtained from a commercial source and the full amino acid sequence of the A1, A2, and the B chains is well established (35). It is conceivable that during the periplasmic expression of CTx the A1 and A2 chains are partially trimmed at either or both termini, thus generating smaller polypeptide chains, similar to that reported previously for the related SLTx A chain (37).

FIGURE 3.

Schematic representation of the CTx holotoxin, with the amino acid residues at the amino (N) and carboxyl (C) termini of the A1, A2, and B chains indicated. Shown is the complete CTx holotoxin (A), along with the two predicted cleavage products (B and C).

TOFMS for the AB₅ holotoxin

ESI-TOFMS was used to try to observe the complete noncovalent CTx AB₅ complex in the gas phase. Abundant ions were observed for the B subunit monomer and the B₅ pentamer, with less abundant ions observed for the AB₅ holotoxin ions at pH 7.0 (Fig. 4 A). The base peak in the mass spectrum corresponds to the monomer B⁷⁺. The partial mass spectrum has been expanded in the region m/z 3000–5000 (Fig. 4 B). The abundances of the B₅ pentamer and the AB₅ holotoxin ions remained in fairly constant ratio over the pH range 4.0–7.0 under the condition specified (partial mass spectrum obtained at pH ∼ 4.0; Fig. 4 C). The abundances of the monomer and pentamer ions observed did not remain constant over the pH range. At pH 5.0 the monomer B⁷⁺ ion had a relative abundance of ∼93% of the base peak corresponding to the pentamer ion. At pH 4.0, the monomer B⁷⁺ ion was observed in the mass spectrum as the base peak and the pentamer ion had a relative abundance of ∼64% to the monomer B⁷⁺ ion. The observed charged states for the AB₅ holotoxin extend from 21⁺ to 15⁺. It is of interest to note that both the AB₅ and B₅ ions were observed throughout the pH range examined, suggesting that the hexameric AB₅ CTx assembly and the pentameric B₅ CTx B assembly were both stable at acidic pH. This is in marked contrast to findings with the structurally related AB₅ SLTx, where the cell-binding B₅ pentamer was less stable at acidic pH than the complete AB₅ holotoxin (37). We therefore suggest that the presence of the A2 portion of CTx A chain is not required to stabilize the CTx B pentamer in the gas phase, as has been postulated for SLTx (37,41,42), but rather that stability arises from the inherent properties of the CTx-B₅. This is consistent with the known stability of CTx B in aqueous solution (6,7,43) where the presence of the A chain had little effect on the stability of the B pentamer.

FIGURE 4.

ESI-ToF mass spectra obtained for the AB₅ cholera holotoxin buffered with ammonium acetate at pH ∼ 7.0 (A). Also shown is the noncovalent assembly at pH 7.0 expanded in the region m/z 3000–4500 (B) and the noncovalent assembly at pH 4.0 expanded in the region m/z 3000–4500 (C).

The heterogeneity of the CTx A chain observed in the triple quadrupole MS was further examined by TOFMS under conditions that maintained the noncovalently associated holotoxin intact in the gas phase. Transformation of the spectrum in the region of m/z 3000–5000, which corresponds to the ions observed for the AB₅ holotoxin, gave suggested molecular mass of 84612.8 Da (calculated 84608.1 Da) and 84932.4 Da (calculated 84932.3 Da) for the heterogeneous AB₅ holotoxin. The results obtained demonstrate that the masses are in good agreement with those expected for the hexameric complex, calculated previously from the denatured complex (Fig. 2).

We have demonstrated that the complete noncovalent assembly of the CTx holotoxin can be detected by ESI-TOFMS over a range of physiological pHs. To our knowledge, this is the first time the CTx holotoxin assembly has been observed intact in the gas phase by MS. The increased sensitivity afforded by TOF mass analyzers for observing such AB₅ assemblies intact in the gas phase supports and extends our previous work (37). The high degree of sensitivity also allowed for the accurate detection of differentially trimmed species of the A-chain component of the holotoxin preparation. Over 2 min acquisition time at a flow rate of 4 μL/min, 8 μL of 8 pmol/μL of AB₅ solution was consumed to generate the mass spectra shown.

The TOF instrument is not capable of MS/MS experiments and further investigative studies were performed in a triple quadrupole instrument. The CTx AB₅ ions observed could not be subjected to MS/MS experiments because of the upper mass range of 4000 of the triple quadrupole instrument. Therefore, further MS/MS experiments were carried out solely on the CTx B₅ pentamer.

Triple quadrupole MS for the homopentamer B₅

The CTx pentamer B₅ consists of five identical subunits that arrange into a pentameric ring structure (Fig. 1 B). Each subunit has a molecular mass of 11.6 kDa and is composed of 103 amino acids. Initially, the mass spectrum was obtained for the noncovalent complex of CTx B₅ buffered with ammonium acetate at pH ∼ 6.6 (Fig. 5). Ions corresponding to the CTx B₅ homopentamer were observed at a range of pH from ∼4.0–7.0 (data not shown). The mass spectrum was obtained with a declustering voltage of 60 V. The base peak in the mass spectrum corresponds to The dominant signal corresponds to the intact homopentamer (B₅), with observed charge states extending from 18⁺ to 15⁺. The pattern of abundance of these ions is identical to the multiply charged species observed with the SLTx B₅ ions, which extended from 15+ to 12+ (37). This presumably reflects the overall similarity in tertiary and quaternary structure between the two toxins. The mass spectrum also shows monomer ions (B) with observed charge states from 8⁺ to 5⁺. The mass spectrum demonstrates that the homopentamer B₅, the dominant quaternary form of the B subunit protein in solution (44–46), remains intact during the ESI process. Furthermore, the B₅ complex remained intact over a range a pH. Over 3 min acquisition time, at a flow rate of 4 μL/min, 12 μL of 4.5 pmol/μL of B₅ solution was consumed to generate the mass spectra shown.

FIGURE 5.

ESI-triple quadrupole mass spectrum of the noncovalent complex of CTx-B₅ at pH 6.6.

It has been reported that low pH (∼5.0), similar to that found within the endosome compartment of intoxicated target cells, causes a conformational change within the CTx B₅s (47,48). However, this conformational change did not promote bound CTx B to dissociate from the GM1 receptor, but rather caused specific changes that affected the structure of the interface between the protein and the membrane to which CTx B was bound (47). If we presume that under our experimental conditions CTx B undergoes a similar structural alteration, we would not predict a destabilization of the B₅ pentameric assembly in the gas phase. Given that the GM1 binding site, as revealed by the crystal structure, is formed between two adjacent monomers, our data are completely consistent with this, insofar as low pH does not alter the binding of CTx to GM1 (47) or, as reported here, does not destabilize the noncovalent CTx B-pentamer.

Triple quadrupole MS/MS of the homopentamer B₅

MS/MS was then used to investigate the dissociation pathway of the homopentamer ion (Fig. 6). The two dominant ions in the mass spectrum of the homopentamer correspond to and (Fig. 5). The ion was selected as a precursor ion to undergo CID in the collision cell of the instrument, the collision energy being varied (15–45 eV (E_LAB = 255–765 eV)) to obtain intense fragment ions (product ions). An increase in the abundance of product ions was observed as the collision energy was increased (Fig. 6 B). The multiply protonated pentamer species selected for CID dissociates almost entirely via the loss of one subunit (Fig. 6 B) to form primarily abundant monomer ions having a range of charges. A very low abundant dimer ion is also observed. The dissociation formed monomer, dimer, trimer, and tetramer ions, but no peaks which could be unequivocally assigned to trimer ions were observed. These ions could still be formed, however, as the ions and cannot be distinguished from the ions B⁺, B²⁺, and B³⁺, and and and respectively, to which some of the peaks have been assigned. However, not all of the ions that are predicted to form were detected (Fig. 6 C). This was either due to them being present in very low abundance or due to the m/z ratio of the ion being beyond the upper m/z range of the instrument, 4000, as discussed above. This is important since the pentamer ion selected for dissociation does not only form monomer ions with a range of charges, it must form tetramer ions with a range of charges also. The tetramer ions formed though, for example will unfortunately not be detected due to the limited m/z range of the instrument.

FIGURE 6.

ESI-Triple quadrupole MS/MS mass spectra of the noncovalent complex of CT-B₅. Shown is the dissociation obtained at a collision energy of 15 eV (A) and 45 eV (B) for the precursor ion Shown is the possible dissociation pathway for the ion (C).

The dissociation of 50% of the CTx precursor ion required a collision energy of 40 eV, (E_LAB 680 eV) ∼2.4-fold higher than that seen for the SLTx precursor ions (Fig. 7) and (37), which both showed 50% dissociation of the precursor ions at ∼20 eV (E_LAB 280 eV).

FIGURE 7.

Plot of the relative ion intensity for the SLTx ion (open symbols) and the CTx ion (solid symbols) versus the collision energy (eV) applied to the ions. The amplitudes of the energy of the two curves at 50% of the intensity provide a measure of the relative gas phase stability of the complexes.

A rough comparison of the relative stability of the two complexes may be gained from a comparison of the relative collision energies required to reduce the intensities of the two precursor ions, (CTx) and (SLTx), to 50% of their initial abundances at similar collision gas pressures. An exact comparison cannot be made since the ions carry different numbers of charges and the distribution of products is unknown in each case. For the CTx complex, a 50% reduction in the abundance of the ion was observed when a potential difference of 40 V was applied between the source and the RF-only collision cell, corresponding to a laboratory collision energy, E_LAB, of 17 × 40 = 680 eV. For the SLTx complex, a similar reduction in the abundance of the ion was observed with a potential difference of 20 V, corresponding to a laboratory collision energy, E_LAB, of 14 × 20 = 280 eV. The approximate masses of the two complexes are 58 kDa and 38 kDa, respectively, and the collision gas was argon, mass 40 Da. The amount of energy available to effect dissociation, the center-of-mass collision energy E_COM, is given by E_COM = E_LAB × (m_Ar/(m_Ar + m_ion)) × 96.5 kJ mol⁻¹, where E_LAB is the ion kinetic energy in the laboratory frame of reference, m_Ar is the mass of argon the collision cell gas, and m_ion is the mass of the precursor ion. This leads to values of ∼45 kJ mol⁻¹ and 28 kJ mol⁻¹ for the CTx and SLTx complexes, respectively, implying that ∼1.5 times as much energy is required to effect the dissociation of the CTx pentamer compared to the SLTx B complex.

This difference in pentamer stability is also seen in aqueous solution (44,46,49). Despite these differences, a high degree of structural stability exists in both the CTx and SLTx B₅ pentamers at physiological pH (37). This has been attributed to the large number of intersubunit interactions that occur between monomers (6,50). The amount of surface area buried during pentamer assembly differs between the two toxins, with CTx B having a greater buried surface area (2700 A²) than SLTx B (>1269 A² < 2700 A²) (6,51). The degree of buried surface area correlates well with the measured stabilities of the two pentamers (44), implying that additional interactions must be present within the CTx B monomer-monomer interface to account for its increased stability. We therefore compared the monomer interface region in the two different pentameric proteins. This revealed that the CTx B monomer interface has an additional α-helix which lies in front of the continuous β-sheet that forms the majority of the CTx B monomer interactions (Fig. 8 A) and this region of secondary structure contains residues that interact with residues that lie in the adjacent monomer. The SLTx B monomer interface region, however, possesses only the continuous β-sheet structure and not the α-helix (Fig. 8 B). In silico energy calculations were performed using the crystal structures of the CTx B (PDB code 1FGB) and SLTx B (PBD code 1BOV) pentamers to assess the theoretical interactions that exist between the additional helix present in the CTx B structure that can interact with the adjacent monomer (Table 1). The comparison of the pentameric structures provides an insight into the apparent stabilization of CTx B relative to SLTx B. The additional N-terminal helix shown in the CTx B structure is involved in a number of extra interactions between monomers such that the intermonomer interaction energy is ∼2.4 times greater than between the monomers of SLTx B. This value is in agreement with that determined experimentally for the increase in energy required to dissociate the two complexes. This revealed that the additional interactions in the CTx B pentamer would most likely result in an increased degree of stability of ∼236 kJ mol⁻¹ to this molecule compared to the SLTx B pentamer. Although the gas phase experimentally determined values are much lower than the theoretically calculated values of 405.7 and 169.5 kJ mol⁻¹, respectively, for monomer-monomer interactions (Table 1), this is to be expected because of the contribution of coulombic repulsion in the fragmenting of the highly charged complexes. The gas phase stability calculation assumes that a single collision has occurred to the precursor ion; however as is the case for large complexes studied here, to reduce the precursor ion intensity by 50%, multiple collisions will occur and contribute to the E_COM value. As gas phase complexes are desolvated before mass analysis, the hydrogen bonding and electrostatic interactions that survive ESI are possibly accountable for maintaining the quaternary structure of the desolvated complexes, whereas the hydrophobic interactions may be lost or partially lost in the gas phase. As such, complete correlation of gas phase and solution phase binding energies would not be expected. Rather, the experimental and theoretical results above show a similar trend, implying that the dominant interactions stabilizing the quaternary structure are maintained in the gas phase and can provide a rough comparison to those in solution. The overwhelming majority of the difference in energy can be accounted for by the presence of a strong hydrogen-bond interaction between E11 in the first monomer with R35 in the second (Fig. 8 C). Also there is a significant area of hydrophobic contact between the N-terminal underside of the additional helix and the top of the nearby sheet from the neighboring monomer (Fig. 8 C). It is important to note that the N-terminal α-helix is held in place by an intramolecular disulphide bridge (C9–C86), which therefore helps to maintain a nonplastic surface for the monomers to interact. We propose that although the interface regions between adjacent monomers has a larger surface area in the CTx B pentamer, the additional interactions present in the CTx B helical region (residues 1–14) may provide the increase in structural stability observed in both the gas and aqueous phase. To further investigate these structural differences in the gas phase, a more detailed comparative study is now required between different B₅ pentamers.

FIGURE 8.

Crystal structures of adjacent monomers within the CTx B and SLTx B pentamers. Monomers from CTx B (A) and SLTx B (B) viewed either sideways (left) showing the interface region (dashed line) between monomers or from the top (right) of the molecule. Structures colored to indicate secondary structure (α-helix in red and β-sheet in yellow). Proposed intermonomer interactions present in CTx B. The chain on the left is chain E; the one on the right is chain D. Chain E is colored differently from D to distinguish between the two. The critical residues in the intermonomer interactions are picked out. E11/R35 are shown in standard atomic colors. Residues involved in hydrophobic interactions are colored pink. The hydrogen bond between E11 and R35 is shown, as is the disulphide bridge. Image was made using ICM.

TABLE 1.

Monomer-monomer interaction energies for the CTx B and SLTx B pentamers^*

Protein	Bond†	Bond energy‡	Protein	Bond	Bond energy
CTx B	vdw	−230.3	SLTx B	vdw	−123.9
(1fgb)	hb	−76.6	(1bov)	hb	−5.4
	el	−98.8		el	−40.2
	total§	−405.7		total	−169.5
		(−236.2)¶
		Gas phase stability‖			Gas phase stability
CTx B		45.1	SLTx B		28.1
		(17.0)**

^*Intersubunit interactions calculated from the x-ray structures (PDB codes) and experimental gas phase dissociation data.

^†Bond type: vdw; van der Waals contact contribution (i.e., hydrophobic); hb, hydrogen bond contribution; and el, electrostatic contribution (but not including H-bond).

^‡Bond energy calculated using Molsoft ICM program (kJ mol⁻¹).

^§Total calculated energies associated with the monomer-monomer interactions.

^¶Calculated energy difference between CTx B and SLTx B.

^‖Experimental energy derived from the center of mass collision energy (kJ mol⁻¹).

^**Experimental energy difference between CTx B and SLTx B.

CONCLUSIONS

As with all types of gas phase experiments, care must be taken in the interpretation of the data, as normally the proteins studied reside and function within an aqueous environment. We believe that the data obtained in this study, however, are highly relevant and informative, particularly since the pH conditions used were near-physiological. As in our previous study (37), data were obtained over a range of pH conditions that correlates well with those encountered by the toxin within the intestinal lumen during infection by V. cholerae. Data for both the CTx AB₅ holotoxin and the B₅ pentamer have demonstrated that the noncovalent interactions within the toxin assemblies are maintained, and observable, under both acidic and neutral conditions in the gas phase. Furthermore, unlike the SLTx B₅ pentamer, the CTx B₅ pentamer is stable at acidic pH, indicating that although the two related pentamers share a similar fold at their β-sheet monomer-monomer interfaces (50), additional interactions must be present within the CTx B pentamer to account for its increased stability. Since the initial stages of CTx mediated cellular signaling are dependent on receptor oligomerization after binding of the pentameric B chains, a rapid, sensitive technique to examine the quaternary structure of both native and any genetically engineered protein variants of CTx will be beneficial to further research performed on this toxin.