Capture of Sulfur Mustard by Pillar[5]arene: From Host-Guest Complexation to Efficient Adsorption Using Nonporous Adaptive Crystals.

Sulfur mustard (SM) has been the most frequently used chemical warfare agent. Here, we present the efficient containment of SM and its simulants by per-ethylated pillar[5]arene (EtP5). EtP5 exhibited strong binding abilities toward SM and its simulants not only in solution but also in the solid state. The association constant (Ka) between SM and EtP5 was determined as (6.2 ± 0.6) × 103 M-1 in o-xylene-d10. Single crystal structure of SM@EtP5 showed that a 1:1 inclusion complex was formed, which was driven by multiple C-H···π/Cl/S and S···π interactions. In addition, activated crystal materials of EtP5 (EtP5α) could effectively adsorb SM simulants at solid-vapor phase; powder X-ray diffraction patterns and host-guest crystal structures indicated that the uptake process triggered a solid-state structural transformation. More interestingly, the captured guest molecules could be stably contained in EtP5α for at least 6 months in air at room temperature.

Introduction

Sulfur mustard (bis(2-chloroethyl)sulfide, SM) is a well-known chemical warfare agent that was first used in World War I (Evison et al., 2001; Szinicz, 2005; Wattana and Bey, 2009). It is referred to several other names (e.g., H, HD, lost, Yperite, and yellow cross liquid) and is the most frequently used chemical warfare agent based on its ease of preparation and its ability to be stored in large quantities. To date, the number of casualties due to SM is greater than the total number of casualties from all other chemical weapons (Szinicz, 2005). Consequently, SM has been called the “King of war gases.” As a powerful blistering agent or vesicant, SM can cause serious damage to eyes, skin, and lungs. Moreover, as an active alkylating agent, it can form a highly reactive three-membered cationic sulfonium intermediate that reacts with DNA bases (Fruton et al., 1946). Despite the development of more toxic chemical weapons (e.g., nerve agents) after SM, the latter is still regarded as one of the most efficient chemical agents in modern warfare (e.g., in warfare of Egypt against Yemen [1960s], Iraq against Iran, Kurds [1980s], and Sudan against insurgents [1990s]) and terrorist threats/attacks.

Therefore, considerable efforts have been devoted to exploring the complexation (Wang et al., 2013), sensing/detection (Jang et al., 2015; Kumar and Anslyn, 2013), and detoxification (Liu et al., 2017; Bobbitt et al., 2017; Decoste and Peterson, 2014) of SM or in most cases its structurally similar molecules. Additionally, capture of other toxic sulfur compounds such as hydrogen sulfide (H2S) and sulfur dioxide (SO2) by metal-organic frameworks (MOFs), zeolites, and carbon-based materials has been demonstrated (Shah et al., 2017; Tchalala et al., 2019; Islamoglu et al., 2020). It is significantly important to develop effective absorption materials for protection from SM damage. However, simple and straightforward physical adsorption materials for SM have been rarely reported. As is well known, it is a powerful strategy to utilize host-guest encapsulation to construct adsorption materials due to the preorganized cavities and multivalent binding sites of the molecular containers (Alsbaiee et al., 2016; Mei et al., 2019; Yu et al., 2012; Wang et al., 2009; Kim et al., 2014; Chen et al., 2013; Schneider et al., 2016; Ajami and Rebek, 2013). To the best of our knowledge, strong interactions of SM, or its stimulants, by macrocyclic hosts have not been demonstrated, because traditional macrocycles have no efficient recognition sites to interact with the small and neutral SM.

As detailed below, we report surprisingly strong interactions of SM by per-ethylated pillar[5]arene (EtP5) and efficient containment of SM and its simulants by activated crystalline materials of EtP5 (EtP5α). Pillar[n]arenes have been regarded as an important family of supramolecular macrocycles (Ogoshi et al., 2008, 2016; Cao et al., 2009; Xue et al., 2012) and have found applications in chemistry (Ke et al., 2013; Li et al., 2014, 2015; Li, 2014; Zhang et al., 2018, 2020; Guo et al., 2018; Kaizerman-Kane et al., 2019), biomedicine (Joseph et al., 2016; Chang et al., 2014; Duan et al., 2013; Li et al., 2017; Zhu et al., 2019; Chen et al., 2019; Si et al., 2015), and materials science (Strutt et al., 2012; Tan et al., 2014; Wang et al., 2016, 2018; Sun et al., 2018; Muhammed et al., 2019; Jie et al., 2018a, 2018b; Song et al., 2018; Ogoshi et al., 2015; Li et al., 2019; Zhou et al., 2019). Host-guest complexation in solution and in the solid state, as well as capture of SM simulants vapor, were comprehensively investigated. Particularly, the captured SM simulants by EtP5α cannot be released for at least 6 months.

Recently, the adsorption and photocatalytic oxidation of SM simulants by MOFs have been well demonstrated (Ma et al., 2019; Cao et al., 2019; Lee et al., 2020; Giannakoudakis and Bandosz, 2020). The present system based on organic macrocycles has the following characteristics (the advantages of pillar[5]arene over MOF are summarized in Table S1): (1) easy accessibility: EtP5 can be prepared through a single-step reaction in high yield of more than 70%, and all reagents are cheap and commercially available; (2) good stability: crystalline materials of EtP5 are stable to moisture; (3) high complexation ability and adsorption stability: SM and its simulants can be quantitatively prisoned in pillar[5]arene containers for quite a long time; (4) convenient modification: mono-, di-, tetra-, and per-substituted pillar[5]arenes can be conveniently obtained. These characteristics would make the present capture of SM find potential applications in degradation and protection materials.

Results

Host-Guest Binding in Solution

Caution

SM is a reactive alkylating and cytotoxic agent. The experiments of host-guest complexation studies were carefully carried out in a fume hood with adequate protection by experienced personnel (more details see Transparent Methods in the Supplemental Information).

First, 1H NMR studies were used to investigate host-guest binding behavior of EtP5 and SM in solution. Figure 2 shows the 1H NMR spectra of SM that were obtained in the absence and presence of one equivalent (eq.) of EtP5 in o-xylene-d10. Proton Hb gives a very large upfield shift of −0.42 ppm and Ha shows so remarkable broadening that its signal could not be observed. At the same time, protons H1,3,4 of EtP5 exhibit downfield shifts (Δδ = 0.26–0.30 ppm) as a consequence of de-shielding effects. These observations indicate an unambiguous binding event.

Figure 2

1H NMR Spectra (400 MHz, 298 K, D2O) at 5.0 mM in o-xylene-d10

(A) EtP5.

(B) SM + EtP5.

(C) SM.

Five SM simulants, S1‒S5 (Figure 1), were also examined as guest molecules in order to systematically study host-guest inclusion behavior. Among these simulants, 2-chloroethyl ethyl sulfide (S1), known as “half mustard,” is also poisonous. S2 and S4 have similar structures to SM, with the middle S atom of SM changed to O and CH2, respectively. S3 and S5 are monofunctional analogs of S2 and S4. Upon addition of EtP5, the guest protons appear similar to NMR changes, i.e., remarkable upfield shifts and broadening (Figures S1–S5), supporting the formation of host-guest complexes. For comparison purpose, the interaction between per-ethylated pillar[6]arene (EtP6) and S2 was also evaluated. No obvious complexation was detected (Figure S6), which is reasonable that the guest is too small in comparison with EtP6's cavity.

Figure 1

Schematic Illustration of Compounds Used in This Study

(A) Structures of EtP5 host.

(B) Structures of SM and its simulants (S1–S5).

The binding events were further validated by two-dimensional nuclear Overhauser effect spectroscopy (2D NOESY) experiments. From the NOESY spectra of S1@EtP5, strong correlations were observed between all the guest protons Ha–d and the aromatic protons (H1) and ethyl protons (H3) of the host (Figure S7). Similarly, the NOESY spectra of S2@EtP5 exhibited unequivocal correlation peaks between the methylene protons Ha,b of S2 and the aromatic protons (H1) and ethyl protons (H3) of EtP5 (Figure S8).

Association constants (Ka) of certain host-guest pairs were determined by employing 1H NMR titration methods (Table 1 and Figure S9). A molar ratio plot for EtP5 and SM based on 1H NMR data indicates 1:1 binding stoichiometry (Figure S10). The Ka value of EtP5 toward SM [(6.2 ± 0.6) × 103 M−1] is 4.8 times larger than that for S2 [(1.3 ± 0.1) × 103 M−1], yet it is 2.9 times lower than that for S4 [(1.8 ± 0.3) × 104 M−1]. The middle methylene of the S4@EtP5 complex can mediate additional C‒H···π interactions, resulting in strong host-guest interactions. The binding affinities of SM, S2, and S4 containing two Cl atoms were found to be significantly greater than those for their monofunctional analogs, S1, S3, and S5 (Table 1). This is because the cooperative hydrogen bonds and dispersion forces between the two ends of the guest and the two cavity portals of the macrocycle enhance the host-guest binding affinity.

Table 1

Association Constants (Ka) Corresponding to Interactions between SM and Its Simulants with EtP5

Host	Guest	Ka (M−1)a
EtP5	SM	(6.2 ± 0.6) × 103
EtP5	S1	(2.9 ± 0.3) × 102
EtP5	S2	(1.3 ± 0.1) ×103
EtP5	S3	67 ± 6
EtP5	S4	(1.8 ± 0.3) ×104
EtP5	S5	(7.9 ± 0.6) ×102
EtP6	S2	b

Ka values were determined in o-xylene-d10 by using 1H NMR titration methods.

Host-guest complexation was not observed.

Crystallographic Investigations

Five host-guest crystal structures of SM simulants, S1@EtP5, S2@EtP5, S3@EtP5, S4@EtP5 and S5@EtP5, and particularly, an X-ray crystal of the real chemical warfare agent, SM@EtP5, were successfully obtained. Details of the crystals are summarized in Table S2. As shown in Figure 3A‒3F, the guests are encapsulated at the center of the EtP5 cavity, forming the 1:1 host-guest complexes. In the crystal structure of SM@EtP5 (Figure 3A), there are four C–H···π interactions between the methylenes of SM and the benzenes of EtP5, with C–H···ring distances of 2.72–3.01 Å (Figure 3G). Only one of the end chlorines forms five C–H···Cl hydrogen bonds with the host portal ethyls (Figure 3H), thereby suggesting that SM is a bit longer compared with the host height. This observation is consistent with our previously reported result that linear ditopic guests with four methylene (–CH2–) linkers are suitable axles for pillar[5]arenes (Li et al., 2010). C–H···S hydrogen bonds were also observed between the ethyls of EtP5 and the sulfur atom of the guest (Figure 3I). More interestingly, there exist triple sulfur–arene interactions between the S atom of SM and the benzenes of EtP5, with S···ring distances of 3.95–4.50 Å (Figure 3J) (Meyer et al., 2003). In the crystal structures of S1–S5@EtP5 (Figures 3B‒3F), guests are encapsulated in the center of the EtP5 cavity by C–H···π/Cl/S interactions for S1@EtP5 (Figure S11A), C–H···O/Cl interactions for S2@EtP5 (Figure S11B), C–H···π/Cl/O interactions for S3@EtP5 (Figure S11C), and C–H···π/Cl interactions for S4@EtP5 and S5@EtP5 (Figures S11D and S11E). These results indicated the formation of inclusion complexes with high stabilities in the solid state, where multiple C–H···π and hydrogen bonding interactions are dominant driving forces.

Figure 3

Single Crystal Structures of Host-Guest Complexes

(A) SM@EtP5. (B) S1@EtP5. (C) S2@EtP5. (D) S3@EtP5. (E) S4@EtP5. (F) S5@EtP5. Chlorine is shown in green, sulfur is shown in yellow, and oxygen is shown in red. Noncovalent bonding parameters of SM@EtP5: Hydrogen bond distances (Å) of C–H···π (G), C–H···Cl (H), C–H···S (I), and S···π (J). See Figure S11 and Table S2 and Data S1.

Vapor Adsorption

Based on the above binding investigation in solution and in the solid state, we wonder whether we can realize the adsorption of SM and its simulants by pillararene-based crystalline materials (EtP5α) (Figure S12). Previously, Ogoshi and Huang et al. (Ogoshi et al., 2015; Jie et al., 2018a, 2018b) reported the adsorption and separation behavior of crystalline pillar[n]arenes. This novel class of materials was termed as nonporous adaptive crystals (NACs) (Jie et al., 2018a, 2018b). Very recently, NACs of other macrocycles such as biphen[n]arene (Wang et al., 2019), hybrid[3]arene (Zhou et al., 2020), leaning towerarene (Wu et al., 2020), geminiarene (Wu and Yang, 2019), and tiararenes (Yang et al., 2020) have also been reported, showing interesting adsorption and separation properties. Notice that, in our vapor adsorption tests, SM simulants were utilized because no effective protection was available during the experiments for testing the real SM.

Figure 4 shows the time-dependent solid-vapor sorption results of EtP5α toward SM simulants. The uptake amount of S1‒S5 in EtP5α increased over time. It was found that the similar adsorption amount of each simulant was observed at equilibrium with the guest/host ratio reaching 1.0–1.2/1 eq. according to 1H NMR results (Figures S13‒S17), indicating SM simulants are contained in the bulk of EtP5α crystals. Thermogravimetric analysis (TGA) measurements further confirmed the quantitative adsorption (Figure S18). EtP5α showed no apparent weight loss after heating to 200°C, demonstrating its stability (Figure S18A). Meanwhile, S1@EtP5 had a mass loss of 12.4% below 200°C. The weight loss was calculated to be 1.0 S1 molecule per EtP5 molecule (mol/EtP5), which is typical for the vapor adsorption results. Intriguingly, we found that the adsorbed guest molecules could not be released by EtP5α upon exposing in air at room temperature for at least 6 months (Figures 4F and S19‒S23), indicating the activated crystal materials of EtP5α have ultra-stable containment ability for SM simulants. What's more, SM simulants could be totally removed under vacuum at 80°C (Figure S24). There was no loss of uptake capacity of EtP5α for adsorption of SM simulants after multiple cycles (Figure S25). In addition, the energies of interactions for the complexes of SM@EtP5 and S1@EtP5 in the crystal state were calculated by molecular mechanics calculations (details in Supplemental Information). The calculated binding energy between EtP5 and SM was 144.6 kJ mol−1, which is higher than that for S1@EtP5 (125.7 kJ mol−1). These results are in accord with the complexation selectivity in solution.

Figure 4

Time-Dependent Vapor Sorption Plot of SM Simulants Vapor by EtP5α at 25°C

(A) S1.

(B) S2.

(C) S3.

(D) S4.

(E) S5.

(F) Recorded residual guest/host ratio of adsorbed S1‒S5 vapor by EtP5α after exposure in air for 6 months.

In order to investigate the mechanism upon uptake of SM simulants vapor by EtP5α (Figure 5G), powder X-ray diffraction (PXRD) experiments were carried out. The PXRD patterns of EtP5α after adsorption of SM simulants were completely different from the pattern for EtP5α (Figure 5H) (Figure S26), indicating that the capture of SM simulants induces the structural transformation of the host crystals. Meanwhile, the PXRD patterns of EtP5α after adsorption of SM simulants were almost the same with each other as well as are in good agreement with that simulated from S1‒S5@EtP5 crystals (Figures 5I‒5M and S26), suggesting that nonporous adaptive crystals of EtP5α transform to the same structure from EtP5α to S1‒S5@EtP5. As can be seen from Figures 5B‒5F, each EtP5 molecule was perpendicular to form herringbone structure with a body-to-window packing mode in S1‒S5@EtP5, which is different from the SM@EtP5 with a window-to-window packing mode (Figure 5A). In addition, it is noted that the structure of S1@EtP5 could transform back to the original structure of EtP5α by removal of S1 under vacuum at 80°C after 12 h (Figure S27).

Figure 5

Crystal Packing Structures and PXRD Patterns

(A–F) Crystal structures of (A) SM@EtP5, (B) S1@EtP5, (C) S2@EtP5, (D) S3@EtP5, (E) S4@EtP5, and (F) S5@EtP5.

(G) The structure of EtP5α (Jie et al., 2018a, 2018b).

(H) Simulated PXRD pattern from the single crystal structure of SM@EtP5.

(I–M) The comparisons of PXRD patterns. Top: simulated PXRD patterns from single crystal structure of S1‒S5@EtP5; bottom: PXRD patterns of EtP5α after uptake of S1‒S5 vapor.

Discussion

In summary, we have demonstrated efficient containment of SM and its simulants by pillar[5]arene, which is the first example of strong interactions of SM by macrocyclic receptors. 1H NMR, 2D NOESY, and X-ray single crystal analysis show that strong binding events occur not only in solution but also in the solid state, which are driven by multiple C–H···π/Cl/S/O hydrogen bonding interactions. In addition, adaptive EtP5α crystals can be a great adsorbent to efficiently and quantitatively adsorb SM simulants with the guest-induced solid-state crystal structure transformation and the captured SM simulants could be contained in the materials for at least half a year. Considering the easy and cheap accessibility, convenient modification, and effective adsorption properties of pillar[5]arenes, the interesting encapsulation of SM holds great potential for the sensing, detoxification, and protection materials for this chemical warfare agent. Endeavors to explore these possibilities are currently underway.

Limitations of the Study

As an effective protective material, crystalline ethylated pillar[5]arene cannot be used for degradation of sulfur mustard. Further investigations would focus on detoxifying mustard gas by functionalized pillar[5]arene. In addition, the TGA experiment for SM complex and the vapor adsorption test for SM were not determined owing to no effective protection for the highly toxic SM.

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Chunju Li (cjli@shu.edu.cn)

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplemental Information. Additional data related to this paper may be requested from the authors. Supplementary crystallographic data for this paper can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. CCDC numbers are 1831237, 1884850, 1831239, 1884851, 1831240, and 1884852, respectively.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.