1 Introduction

Cholesterol sulfate (CholSO4) is a naturally occurring cholesterol (Chol) analogue widely found in various tissues [1–5]. Its physiologic function is only partially understood. However, CholSO4 is generally known to play roles in stabilizing and modifying the properties of cell membranes. It protects erythrocytes from osmotic lysis [2] and regulates sperm capacitation [3]. A recent study using mass spectrometry has reported CholSO4 as a potential biomarker of human prostate cancer [4]. The structure of CholSO4 is identical to Chol except for its anionic sulfate functionality. Both molecules exert comparable effects in stabilizing and modifying properties of cell membranes [5]. However, presence or absence of the sulfate moiety also exerts dissimilar effects on cell membranes. For example, studies have shown that CholSO4 increases the order of acyl chains of phosphatidylcholine for temperatures higher than the gel-to-liquid crystal transition point, while it decreases the order for temperatures below the phase transition point [6–8]. A simulation study of interactions of 1,2-dipalmitoyl-sn-phosphatidylcholine (DPPC) with CholSO4 and with Chol reveals that the hydrophobic rings of both species occupy similar locations in acyl chain bilayers [5]. The simulation also demonstrates that the head group of CholSO4 is located slightly below the lipid phosphate group, forming more hydrogen bonds with water molecules than is the case for the hydroxyl group of Chol [5]. This leads to stronger interactions between DPPC and CholSO4, which increases the order effect on DPPC acyl chains, but the effect is limited to adjacent DPPC.

The double bond of the oleoyl group of monounsaturated phospholipid, which is the most common unsaturated lipid found in all membranes, is located at the middle (ω-9) of the acyl chain [9]. Previous theoretical studies have indicated that this specific location of the cis double-bond allows a maximum surface area per lipid, thus providing the highest fluidity of the lipid layer [9]. The double-bond at ω-9 also weakens condensing and ordering effects induced by Chol [10, 11]. This effect is proposed to be caused by co-localization of the phospholipid double-bond and the hydrocarbon chain of Chol [9]. In contrast to Chol, interactions of CholSO4 in unsaturated phospholipid layers have not been as thoroughly investigated. As reported in previous studies of a saturated phospholipid membrane, the common hydrophobic rings and tails of CholSO4 and Chol may exhibit similar interactions with acyl chains of unsaturated phospholipids, with the different head groups being the main source of distinctive behavior.

Field-induced droplet ionization mass spectrometry (FIDI-MS) employs a soft ionization technique to sample ions from the surface of microliter liquid droplets to a mass analyzer [12–15]. Utilizing FIDI-MS, several studies of time-resolved heterogeneous reactions at air-liquid interfaces have been reported [14, 15]. For example, the location and orientation of SP-B1-25 (a shortened version of human surfactant protein B) in a lipid surfactant layer has been deduced from experimentally observed reactivity with ozone (O3) using a FIDI-MS methodology [14]. In addition, detailed mechanistic studies of the heterogeneous oxidation of unsaturated phospholipids as well as the alteration of phospholipid compositions resulting from reaction with O3 at the air-liquid interface have also been reported [15].

In this study, we investigate the interfacial reactivity of CholSO4 both alone and in the 1-palmitoyl-2-oleoyl-sn-phosphatidylglycerol (POPG) surfactant layer (Scheme 1) at the air-liquid interface of microliter droplet via heterogeneous oxidation by O3 and computational modeling. CholSO4 is used due to its high ionization efficiency and interesting structural and functional characteristics in a lipid membrane, as well as its high similarity to Chol [16]. Especially, CholSO4 has been reported to be miscible both in liquid condensed and liquid extended layers [16]. The unique capabilities of FIDI-MS provide detailed information regarding reactants, intermediates, and products present in the interfacial layers. First, we probe air-liquid interfacial oxidation of CholSO4 by O3. Then, in order to understand characteristics of the air-liquid interfacial ozonolysis of CholSO4, we have performed solution and solid phase reactions with O3 for comparison purpose. Finally, we investigate the air-liquid interfacial reaction of CholSO4 in the POPG layer with O3. To support our interpretation of the experimental results, molecular dynamics (MD) simulations of CholSO4 mixed with POPG in a monolayer on the surface of water have been performed. The MD simulations correlate well with experimental observations and provide additional insights into the interactions between lipids and water molecules in the interfacial region.

Scheme 1
scheme 1

Structures of CholSO4 and POPG anions investigated in this study

Full size image

2 Experimental

2.1 Chemicals and Reagents

The sodium salt of CholSO4 is purchased from Sigma-Aldrich (St. Louis, MO, USA), 25-[N-[(7-nitro-2-1,3-benzoxadiazol-4-yl)methyl]amino]-27-norcholesterol (25-NBD Chol) and the sodium salt of POPG are purchased from Avanti Polar Lipid (Alabaster, AL, USA) and used without further purification. All solvents (water and methanol) are HPLC grade and purchased from EMD Chemicals Inc. (Gibbstown, NJ, USA).

2.2 Air-Liquid Interfacial Oxidation by O3

The FIDI-MS instrument is based on a design previously described by Grimm and Beauchamp [13]. A stainless steel capillary with ~2 mm o.d. hanging droplet of analyte solution is located between the atmospheric sampling inlet of a mass analyzer (Thermo Finnigan LCQ Deca mass spectrometer) and a parallel plate electrode. The droplet is exposed to O3 for a desired period of time between 0 to 60 s. FIDI sampling is then achieved by applying pulsed voltages of –4 and –2 kV to the parallel plate electrode and supporting capillary, respectively. The FIDI-MS spectra reported in this study are obtained by averaging 10–30 individually acquired spectra from separately prepared droplets. Ion abundances are analyzed by measuring peak areas in FIDI-MS spectra. Approximately 20 ppm O3 is generated using a pencil-style UV calibration lamp (model 6035; Oriel). The ozone concentration is measured spectrophotometrically using an absorption cell with 10 cm path length. The ozone concentration is calculated as ~20 ppm in air with a molar absorption coefficient of 1.15 × 10–17 cm2 molecule–1 in a flow that continually bathes the droplet at ~ 1500 mL min–1. 50 μM CholSO4 or a mixture of 50 μM CholSO4 and 50 μM POPG in 1:1 (by volume) water and methanol feed the droplet source.

2.3 Solution Phase and Solid Phase O3 Reactions

A continuous flow of ~20 ppm O3 in He is bubbled into a 50 μM CholSO4 solution (200 μL) in 1:1 (by volume) water and methanol solvent for 15, 30, 45, and 60 s for solution phase O3 reactions. For solid phase reactions, a continuous flow of ~20 ppm O3 in He is applied to a dried ~1.4 × 10–4 g of CholSO4 film in 20 mL glass vial for 30 min, 2 h, and 12 h. The dried CholSO4 film is prepared by drying 300 μL of 1 mM CholSO4 solution dissolved in 2:1 (by volume) chloroform and ethanol solvent under dry N2. The film is then placed under vacuum overnight. For analysis a sample solution is prepared with a total 50 μM concentration using methanol for electrospray ionization (ESI). Product analysis is performed on a Thermo Scientific LTQ Velos dual ion trap mass spectrometer in negative ion mode.

2.4 Fluorescence Microscopy

The fluorescence labeled CholSO4 (25-NBD CholSO4) is prepared from 25-NBD cholesterol using modified method of Duff [17] by Sandhoff et al. [18]. In brief, 0.355 mg of 25-NBD cholesterol is dissolved 20 μL of 5 mg/mL sulfur trioxide pyridine complex solution in absolute pyridine. After 10 min in room temperature, 2.1 μL of 314.1 mM barium sulfate is added and left for 10 min in room temperature. Then, the sample is incubated in distilled water in the refrigerator (4°C) for 1 h. Lastly, the solution is centrifuged at 10,000 rpm for 10 min at 15°C and stored at −20°C. The purity of the derived 25-NBD CholSO4 is checked using TLC and single spot is found from normal phase chromatography. Fluorescence microscopy observation of air-liquid interface is carried out using fluorescence microscope (Eclipse 80i; Nikon) with a mercury lamp as a light source. Lipid monolayer is prepared on the 300 μL water or water/methanol (1:1 by volume) droplet, which is deposited on the microscope slide with a cavity. The lipid layer composed with the mixture of 10 μM CholSO4 and 10 μM POPG contains 0.5 mol% of 25-NBD cholesterol sulfate.

2.5 Computational Modeling

Molecular dynamics (MD) simulations are performed with the all-atom CHARMM22 force field [19, 20] using the CHARMM package [21]. Flexible TIP3P water potential is used with Hooke’s constants of 900 kcal mol–1 Å–2 for the OH bond and 110 kcal mol–1 rad–2 for the H-O-H angle. Models of CholSO4 and POPG lipid monolayer-water system consist of 48 hexagonally-packed lipids and are simulated at different surface densities of 55, 60, 65, and 70 Å2 per lipid molecule. A wall potential described by the repulsive part of V = ε[2/15(σ/r)9–(σ/r)3] with ε = 0.1521 kcal/mol and σ = 3.1538 Å is employed to prevent water molecules from diffusing out of the box. The force field parameters of CholSO4 are generated based on all-atom Chol parameter set reported by Pitman et al [22]. The modifications are on the head group of CholSO4: the parameters of S and O atoms are directly adopted from methylsulfate parameters [23, 24] and the charge on the linking carbon on the cholesteryl ring is adjusted to the correct total charge of the entire molecule (−1). These systems are composed of a 1:3.4 CholSO4: POPG ratio and generated in two dimensions. The box dimensions of the MD simulations are (55.21 × 55.21 × 59.82 Å) for the 55 Å2/lipid case, (57.67 × 57.67 × 59.82 Å) for the 60 Å2/lipid, (60.02 × 60.02 × 59.82 Å) for the 65 Å2/lipid, and (62.28 × 62.28 × 59.82 Å) for the 70 Å2/lipid. Electrostatic and Lennard-Jones interaction were considered with 12 Å cutoffs and 10 Å tapers. The switching is performed so that the force, not the potential, smoothly decays to zero within this 2 Å tapering region. Each simulation consists of 0.5 ns equilibration at 300 K using Nose-Hoover thermostat NVT MD simulations with a relaxation time of 0.1 ps, followed by 2.0 ns of production NVT simulations for the molecular distribution analysis.

3 Results and Discussion

3.1 Air-Liquid Interfacial Reaction of CholSO4 with O3

The negative ion FIDI-MS spectra for ozonolysis of CholSO4 in a water/methanol (1:1 by volume) droplet are shown in Figure 1a. Prior to O3 application, singly deprotonated CholSO4, observed at m/z 465, is seen as a dominant species in the FIDI-MS spectrum. Highly abundant products resulting from ozonolysis of CholSO4 at the air-liquid interface appear as early as 5 s after exposing the droplet to O3. The products at m/z 481, 497, and 513 correspond to singly (+16), doubly (+32), and triply (+48) oxygenated CholSO4, respectively. Based on the previous studies of ozonolysis of Chol [25–29], singly, doubly, and triply oxygenated products are suggested as epoxide (product I, m/z 481), dicarbonyl (product II, m/z 497), and carbonyl-acid (product III, m/z 513), respectively. Alternatively, the vinyl hydroperoxide may be generated for product III [26, 30]. In addition, what we assume to be hydroxyhydroperoxide (HHP), methoxyhydroperoxide (MHP) products are observed at m/z 531 and 545, respectively [29, 31]. The proposed structures of products and reaction mechanisms are shown in Scheme 2. There are structural isomers and many reaction mechanisms may exit, but only some pathways of the observed products are summarized here. This putative analysis needs further validation using accurate mass measurement, which is beyond the scope of the present study.

Figure 1
figure 1

(a) Air-liquid interfacial reaction of CholSO4 with O3 as a function of time. The negative ion FIDI-MS spectrum of CholSO4 is dominated by the singly deprotonated CholSO4 peak at m/z 465 in the absence of ozone. The products appear after the droplet is exposed to O3 for 5 s. After 15 s of ozone exposure, the FIDI-MS spectrum is dominated by oxidative products. Deprotonated HHP (product IV) and MHP (product V) products are observed at m/z 531 and 545, respectively. Singly (product I), doubly (product II), and triply (product III) oxygenated products are shown at m/z 481, 497, and 513, respectively. The structure of each product is shown in Scheme 2. (b) The relative abundance of the reactant and products of air-liquid interfacial ozonolysis of CholSO4 in the FIDI-MS spectra as a function of reaction time. (c) Plot of the abundance changes of CholSO4 in the ozonolysis at the air-liquid interface. Solid line is the exponential fit to determine the rate constant

Full size image
Scheme 2
scheme 2

Summary of heterogeneous oxidation of cholesterol sulfate by O3 at the air-liquid interface. R is H for water and CH3 for methanol. Proposed structures of products I, II, and III are adopted from Reference [25–29]. Proposed structures of products IV and V are adopted from Reference [29, 31]

Full size image

The relative abundance of the reactant CholSO4 decreases dramatically after 15 s of exposure, and the FIDI-MS spectrum is dominated by ozonolysis products after 30 s. The relative abundance changes of the reactant and products as a function of reaction time are found in Figure 1b. During the ozonolysis reaction, the ozone concentration is assumed to be constant. Then, the reaction rate of CholSO4 is expressed as

$$ - \frac{{d{{[CholS{O_4}]}_{{surf}}}}}{{dt}} = {k_2}{[CholS{O_4}]_{{surf, 0}}} $$
(1)

where k 2 = k 1 [O3] using the pseudo-first order kinetics. Solving Equation (1) gives

$$ \frac{{{{\left[ {CholS{O_4}} \right]}_{{surf}}}}}{{{{\left[ {CholS{O_4}} \right]}_{{surf,{ }0}}}}} = {e^{{ - {k_2}t}}} $$
(2)

The applied ozone concentration is 5 × 1014 molecule cm–3 (20 ppm). As seen in Figure 1c, k 2  = 0.27 s–1 is obtained from CholSO4 abundance in FIDI-MS spectra. Then, second-order rate constant, k 1 , is determined as 5.4 × 10–16 cm3 molecule–1 s–1. The observed CholSO4 ozonolysis rate constant at the air-liquid interface is comparable to the ozonolysis rate constant of 1-oleoyl-2-palmitoyl-sn-phosphatidylcholin (OPPC) monolayer on NaCl (4.5 × 10–16 cm3 molecule–1 s–1) [32]. Our previous study of POPG ozonolysis at the air-liquid interface (7.4 × 10–16 cm3 molecule–1 s–1) using FIDI-MS has also shown a good agreement with these values [15]. The experimentally determined rate constant values are listed in Table 1.

Table 1 Experimentally determined ozonolysis rate constant values of CholSO4 and POPG
Full size table

It is notable that the oxygenated products I, II, and III appear at the beginning of the ozonolysis (O3 exposure for 5 s). However, as the droplet is exposed to O3 for 15 s, a relatively low abundance of these products is observed in the FIDI-MS spectrum (Figure 1b). After 30 s exposure of O3 to the droplet, only a small amount of product III is observed along with highly abundant product IV and product V. The air-liquid interfacial ozonolysis of CholSO4 yields HHP (product IV) and MHP (product V) products as the most abundant products. In order to yield HHP, a primary ozonide (POZ) reacts with a water molecule (Scheme 2) [32, 33]. However, this unstable product converts to an aldehyde product through a proton transfer with a water molecule in the bulk phase [33]. Abundant hydroperoxide products in the FIDI-MS spectra have been reported in our previous study of heterogeneous POPG ozonolysis [15]. These products are observed in the spectra due to low water abundance at the air-liquid interface. Similarly, the abundant HHP product from CholSO4 ozonolysis is attributed to a relatively low water density in the CholSO4 interfacial layer. At the air-water interface, Chol assembles in ordered crystalline structures up to three molecular layers thick [34, 35]. Analysis using X-ray diffraction indicates that the Chol crystalline phase on the liquid surface comprises Chol monohydrate crystallites [35]. It is expected that CholSO4 interacts with water molecules to a greater extent than Chol due to its anionic sulfate head group. In contrast to the behavior of Chol with the hydroxyl group, solvation of the anionic sulfate group may prevent CholSO4 from forming multilayer crystalline hydrates at the air-liquid interface.

No dimeric or trimeric product is observed in the FIDI-MS spectra. The study of Chol ozonolysis in a gas phase aerosol reported bound multimeric products, which are formed through aggregated gas phase clusters [26]. The absence of multimeric product from heterogeneous ozonolysis suggests that CholSO4 molecules do not aggregate and form well-oriented layer structures on the surface of the droplet.

3.2 Solution and Solid Phase Reaction of CholSO4 with O3

Solution and solid phase reactions with O3 have been performed in order to understand the unique characteristics of air-liquid interfacial ozonolysis of CholSO4. O3 is bubbled into the CholSO4 solution and applied to a dried CholSO4 crystal film for solution-phase and solid-phase reactions, respectively. Figure 2 shows ESI-MS spectra of ozonolysis products of CholSO4 after bubbling of 20 ppm O3 into the 200 μL CholSO4 solution (Figure 2a) and after applying 20 ppm O3 to a dried CholSO4 film (140 μg, Figure 2b).

Figure 2
figure 2

(a) ESI-MS spectra of the oxidized products of CholSO4 from solution phase O3 reaction as a function of time. Singly deprotonated CholSO4 is observed at m/z 465 along with deprotonated singly, doubly, and triply oxygenated products at m/z 481, 497, and 513, respectively. HHP and MHP products are observed at m/z 531 and 545, respectively. (b) ESI-MS spectra of the oxidized products of CholSO4 from solid-phase O3 reaction as a function of time. The singly deprotonated CholSO4 is observed at m/z 465. The products at m/z 499 and 513 correspond to diol product and carbonyl-acid product. The ion at m/z 544.5 is unknown cluster ion comprise ~141.8 mass unit monomer and singly charged anion with 119 m/z

Full size image

All products observed from the air-liquid interfacial ozonolysis of CholSO4 (Figure 1a) are observed from the solution phase ozonolysis with different product distribution (Figure 2a). The solution phase ozonolysis of CholSO4 also yields HHP (product IV, m/z 531) and MHP (product V, m/z 545) products as the most abundant products. The significant difference compared with air-liquid interfacial ozonolysis is observed from the relative abundance of three oxygenated products (products I, II, and III). After bubbling of O3 for 60 s, product I (m/z 481) appears as the most abundant product in the ESI-MS spectrum while product III at m/z 513 is shown as the least abundant product among products I, II, and III.

The solid phase reaction yields different products except for triply oxygenated product (product III) at m/z 513 (Figure 2b). Singly and doubly oxygenated products are not observed at m/z 481 and 497, respectively, in the ESI-MS spectrum after applying O3 to a dried CholSO4 film for 12 h. As well, hydroperoxide products (m/z 531 and 545) are not observed. Instead, a diol (+34) product is observed at m/z 499. The absence of product I (epoxide), product II (dicarbonyl), and hydroperoxide (products IV and V) products from ozonolysis of the CholSO4 film implies that formation of these products requires a wet environment.

3.3 The Role of Water Molecules in Ozonolysis of CholSO4

Previous studies of the ozonolysis of Chol have reported that the epoxide product is observed only in a polar solvent [25, 36]. Formation of this product does not occur in an aprotic solvent [25]. The epoxide product is also observed from free-radical peroxidation processes [36]. Once O3 is dissolved in a polar solvent (i.e., water), unstable O3 rapidly forms secondary reactive oxygen species (ROS), including OH radical [37]. This implies that the formation of an epoxide product (product I in Scheme 2) involves either an ozonolysis process aided by polar solvent or oxidation of CholSO4 with a secondary ROS [38]. However, O3 has limited solubility in an aqueous solution, with a very low Henry’s law constant (0.011 M/atm) [39]. This results in a very limited number of ROS in solution to produce the abundant epoxide product. This suggests that interactions of a Criegee intermediate (CI) or a primary ozonide (POZ) with solvent molecules (water or methanol) would be a responsible for the formation of product I (m/z 481 in Figures 1a and 2a) from the ozonolysis of CholSO4. The role of solvent molecules to yield the dicarbonyl product (product II in Scheme 2, m/z 497 in Figure 1a) is unclear, but dicarbonyl product is known to be highly abundant interfacial ozonolysis product of Chol and even suggested as a biomarker for ozone exposure of lung, where constant heterogeneous reactions occur [27, 28]. The carbonyl-acid product (product III) at m/z 513 is the least abundant product from solution phase ozonolysis (Figure 2a) while it is the most abundant product from solid phase ozonolysis (Figure 2b). The formation of a carbonyl-acid product does not require a solvent molecule. Thus, this product is highly abundant with a dry environment and observed in much lower yields in a wet environment.

As seen in Figure 1b, the abundance of the epoxide product (product I, m/z 481) is comparable to the carbonyl-acid product (product III, m/z 513) after the droplet is exposed to O3 for 5 s. However, its abundance decreases dramatically relative to other products after exposure of the droplet to O3 for 15 s. This indicates an important characteristic of the air-liquid interface. As discussed earlier, low water density in the CholSO4 layer is expected at the air-liquid interface. The limited number of water molecules around the double bond of CholSO4 is consumed to form HHP (product IV). This results in the low yield of the epoxide product via ozonolysis of CholSO4 at the air-liquid interface. A similar phenomenon is observed from the previous study of the ozonolysis of POPG at the air-liquid interface [15]. The secondary ozonide of POPG (POPG-SOZ), which forms only under an anhydrous environment, starts building up after the limited water molecules are depleted around the double bond of POPG. The relatively low abundance of the product II is also a consequence of the limited number of water molecules around the double bond of CholSO4. Solution phase ozonolysis of CholSO4 shows a higher abundance of product II (m/z 497) than product III (m/z 513), while only the latter is observed under a dry environment (Figure 2). Once the CI forms from the POZ of CholSO4 (Scheme 2), after the limited water molecules are depleted, the carbonyl-acid product is preferentially formed at the air-liquid interface.

3.4 Air-Liquid Interfacial Reaction of a CholSO4 and POPG Mixture with O3

Negative ion FIDI-MS spectra for the time resolved ozonolysis of an equimolar mixture of CholSO4 and POPG at the air-liquid interface are shown in Figure 3a. Singly deprotonated CholSO4 and POPG, are observed at m/z 465 and 747, respectively. In the FIDI-MS spectrum, POPG exhibits higher abundance than CholSO4 in the spectrum and the abundance of CholSO4 is ~76% of the POPG abundance.

Figure 3
figure 3

(a) Heterogeneous reaction of a 1:1 mixture of CholSO4 and POPG with O3 as a function of time. In the absence of ozone, singly deprotonated POPG and CholSO4 peaks are observed at m/z 747 and 465, respectively. The oxidation products of POPG and CholSO4 are observed after 5 s of O3 exposure. The structure of each product from POPG is shown in Scheme 3. (b) Plot of the abundance changes of CholSO4 (black square) and POPG (red circle) in the ozonolysis at the air-liquid interface. The abundance changes of CholSO4 and POPG when each molecule is present alone on the droplet are shown as solid square and circle, respectively. The abundance changes of POPG is adopted from Ref [15]. The abundance changes of CholSO4 and POPG in a mixture layer are shown as empty square and circle, respectively. Solid lines are the exponential fits to determine the rate constants when each molecule is present alone and dashed lines are the exponential fits when they are in a mixture layer. (c) The relative abundance of the CholSO4 (black solid square) and POPG (red solid circle) during the ozonolysis at the air-liquid interface in the FIDI-MS spectra as a function of reaction time

Full size image

Our recent study of the ozonolysis of POPG at the air-liquid interface has shown that products of POPG appear as early as 5 s after exposing the droplet to 20 ppm O3 [15]. All products of POPG previously observed, including aldehyde (m/z 637), carboxylic acid (m/z 653), peroxic acid (m/z 669), POPG hydroxyhydroperoxide (POPG-HHP, m/z 671), and POPG methoxyhydroperoxide (POPG-MHP, m/z 685), are also observed from the ozonolysis of POPG in the mixture with CholSO4 (Scheme 3). The POPG-SOZ at m/z 795 shows up after exposing droplet to O3 for 15 s before POPG is depleted. In the mixture, only HHP (m/z 531) and MHP (m/z 545) products are observed in significant yield in the FIDI-MS data.

Scheme 3
scheme 3

Summary of the observed products from heterogeneous oxidation of POPG by O3 at the air-liquid interface. R′ is H for water and CH3 for methanol. Product structures are adopted from Reference [15]

Full size image

Ozonolysis rates of CholSO4 and POPG in the mixture decrease compared with when each molecule is present alone (Figure 3b). The observed rate constant values, k 1 , of CholSO4 and POPG are 3.6 × 10–16 cm3 molecule–1 s–1and 5.9 × 10–16 cm3 molecule–1 s–1, respectively, in a mixture (Table 1). The slower reaction rates of both CholSO4 and POPG suggest that CholSO4 is present with POPG as a well-mixed interfacial layer supporting the previous study of its miscibility to liquid extended layers [40]. Our fluorescence microscopy observation of the mixture of CholSO4 (with 0.5 mol% 25-NBD CholSO4) and POPG at the air-liquid interface also supports this showing homogeneous one phase image (Figure S1 in the Supplemental Information). Overall, the POPG ozonolysis rate decreases by 20% and CholSO4 ozonolysis rate decreases by 34% in a mixed layer (Figure 3b) [15]. As discussed earlier, air-liquid interfacial ozonolysis rate constants of both CholSO4 (5.4 × 10–16 cm3 molecule–1 s–1) and POPG (7.4 × 10–16 cm3 molecule–1 s–1) are comparable. In addition, the ozone concentration is assumed to be constant during the ozonolysis. The observed time delay for overall ozonolysis of POPG and CholSO4 implies that double bonds of POPG and CholSO4 are more shielded by additional hydrocarbons from each other at the air-liquid interface. Slightly longer delay for the ozonolysis of CholSO4 indicates that shielding effect of POPG acyl chain to the double bond of CholSO4 is slightly higher compared with the shielding effect of CholSO4 to the POPG double bond. The relative abundance of CholSO4 increases after 30 s of O3 exposure (Figure 3c). Slightly more reactive POPG yields more hydrophilic ozonolysis products, which diffuse into the aqueous droplet [15] increasing relative abundance of CholSO4 in the interfacial surfactant layer.

It is notable that the POPG-SOZ is formed before POPG is depleted in the spectrum. Previous study of the heterogeneous ozonolysis of POPG has shown that POPG-SOZ starts building up after POPG is depleted on the surface of the droplet [15]. The early appearance of POPG-SOZ suggests rapid depletion of the limited number of water molecules in the hydrophobic portion of the lipid layer due to the co-consumption of water molecules with CholSO4.

3.5 Interactions of CholSO4 and POPG Double Bonds with Water Molecules at the Air-Liquid Interface

MD simulations for the mixture monolayer of CholSO4 and POPG on a water box are performed for 2.0 ns. Four different surface densities (55, 60, 65, and 70 Å2/lipid) of CholSO4 and a POPG mixture monolayer are used for the simulations. These surface densities are reported as a reasonable density range of a lipid monolayer at the air-liquid interface [41–43]. The ratio between CholSO4 and POPG in the monolayer is set to be 1:3.4. In order to investigate the effect of CholSO4 abundance in the monolayer, additional MD simulation for the mixture monolayer of CholSO4 and POPG with ratio of 1:1 is also performed for 2.0 ns using the surface density of 60 Å2/lipid.

Figure 4a shows a final snapshot of the MD simulation, when the ratio of CholSO4 to POPG is 1:3.4, for 2.0 ns with the surface density of 60 Å2/lipid as a representative case. Figure 4b shows the atomic density profiles of oxygen atoms of water molecules, sp2 carbons (double bond) of lipid acyl chains along ±Δz, which is the z-direction location relative to the averaged position of the POPG phosphorous atom resulting from the simulation using 1:3.4 CholSO4 and POPG. The interaction between lipid functional groups and water can be estimated from the density of water molecule oxygen atom at the z-direction location of the lipid functional group. The water density around the double bonds of CholSO4 is ~0.0036 atom/Å3. Compared with the bulk phase where water density is ~0.034 atom/Å3, a limited number of water molecules are present around the double bond of CholSO4 to be involved when ozone reacts with the double bond of CholSO4. As discussed earlier, this limited number of water molecules is mainly consumed to form HHP.

Figure 4
figure 4

(a) Final snapshot after 2.0 ns of MD simulation of a mixture of CholSO4 and POPG monolayer at 60 Å2/lipid. CholSO4 (displayed as tube) is shown in olive (tail) and burgundy (head). POPG, water molecules, and sodium ions are shown in gray, blue, and purple, respectively. (b) Atomic density profiles of a mixture of CholSO4 and POPG monolayer systems as a function of Δz, where the averaged position of the POPG phosphorous atom at the air-liquid interface is 0 (negative values are toward the water layer, and positive values are toward the lipid). The lipid surface densities are 55 Å2/lipid, 60 Å2/lipid, 65 Å2/lipid, and 70 Å2/lipid. Black solid lines denote the density profiles of oxygen atoms of water molecules. Red solid lines denote that of sp3 carbons of POPG acyl chains, and blue solid lines denote that of sp2 carbons of POPG acyl chains. Olive and magenta solid lines denote density profiles of sp3 and sp2 carbons of CholSO4, respectively

Full size image

Based on the observed competitive reactivity of CholSO4 and POPG with O3 (Figure 3a), we have suggested that these molecules are co-located at the interface. Our interpretation is further supported by the MD simulations of a mixture monolayer of CholSO4 and POPG. Overall, POPG and CholSO4 show similar medial thickness on the surface of water. In other words, they are co-located in the lipid monolayer. When the lipid monolayer has a 60 Å2/lipid surface density, hydrophobic carbons of both lipids stand up to ~14 Å. Although the overall dimension of POPG exceeds those of CholSO4, the hydrophobic portions of both molecules have a comparable extent (Scheme 1). As seen in Figure 5a, the sulfate group of CholSO4 directly interacts with as many as eight water molecules via ionic-hydrogen bonds. These strong interactions of the sulfate group with water molecules and the rigid ring structures of CholSO4 induce high tilt angle (67°) from the surface of the water. Therefore, CholSO4 and POPG form a well-mixed interfacial layer on the surface of the water. It is noteworthy that compared with the sp2 carbons of POPG, sp2 carbons of CholSO4 are located only ~4.4 Å closer to the water surface (Figure 5b).

Figure 5
figure 5

Final snapshot after 2.0 ns of MD simulation of a mixture of CholSO4 (lime) and POPG (magenta) monolayer at 60 Å2/lipid. (a) Sulfate group of CholSO4 is interacting with eight water molecules. (b) The sp2 carbons of CholSO4 are located ~4.4 Å below sp2 carbons of POPG. The calculated distances between atoms are indicated as Ångström

Full size image

To understand the observed time delay for overall ozonolysis of both POPG and CholSO4 in the mixed lipid layer, we have calculated the number of sp3 carbons from the surface of surfactant layer to the locations of double bonds, from the atomic density profiles. The calculated number of sp3 carbons per sp2 carbon in this region, using a surface density of 60 Å2/lipid, is summarized in Table 2. When only CholSO4 and POPG are present on the surface of water, 10.8 and 15.8 sp3 carbons surround each sp2 carbon, respectively, and are located in the interior of the hydrophobic carbon layer. For each sp2 carbon of CholSO4 and POPG, the number of sp3 carbons in the mixed lipid layer increases to 70.4 and 18.6, respectively, when the ratio of CholSO4 to POPG is 1:3.4. In the case where the amount of CholSO4 is comparable to POPG in the monolayer (i.e., 1:1), our MD simulation shows each sp2 carbon of CholSO4 is surrounded by 31.3 and 29.5 sp3 carbons in the monolayer, respectively (Figure S2 in the Supplemental Information). This strongly indicates that an increasing number of hydrocarbon chains eclipse the double bonds of CholSO4 and POPG as O3 diffuses into the mixed lipid monolayer. As a result, both POPG and CholSO4 ozonolysis is delayed in a mixed interfacial layer, as we observed from the experiments using FIDI-MS (Figure 3).

Table 2 The number of sp3 carbons per sp2 carbon calculated from the surface of surfactant layer to the locations of double bonds using a surface density of 60 Å2/lipid
Full size table

4 Conclusions

We have utilized the FIDI-MS technique to examine the effect of environmental exposures on the surfactant layer at the air-liquid interface. This study provides details for the reaction of CholSO4 with O3 to understand the unique chemistry of this molecule at an air-liquid interface. Time-resolved studies of ozonolysis of CholSO4 at the air-liquid interface reveal that a limited amount of water around double bonds of CholSO4 plays an important role in yielding oxygenated products. The epoxide and dicarbonyl products are observed only when water molecules are present around double bonds of CholSO4. Competitive oxidation of CholSO4 and POPG at the air-liquid interface suggests that both lipids form a mixed interfacial layer when they are present together in a lipid surfactant layer. In a mixed layer, the double bonds of CholSO4 and POPG are more shielded by additional hydrocarbons from each other resulting in a time delay for the ozonolysis of both molecules. MD simulations of a mixed interfacial monolayer of CholSO4 and POPG provide a detailed picture of the interactions between POPG, CholSO4, and water molecules in the interfacial region. In these simulations, the location and orientation of CholSO4 relative to POPG provide a rationalization for the experimental observations.