Abstract
A series of N-monosubstituted and N,N’-disubstituted derivatives of the indolo[3,2-b]carbazole chromophore have been prepared, and their binding affinity for duplex DNA has been evaluated by ultraviolet and fluorescence spectroscopies. It has been found that indolo[3,2-b]carbazoles bearing basic N-alkyl substituents are intercalators that bind DNA with affinities in the micromolar and submicromolar range and a preference for associating with sequences of mixed composition and purine-pyrimidine steps.
Over the past decade, numerous derivatives of the indolo[3,2-b]carbazole chromophore have been prepared in order to explore their photophysical properties and their application as materials for organic electroluminescent devices.1 The precursor to these substances, the chromophore indolo[3,2-b]carbazole, contains two nitrogen atoms on opposing sides of the aromatic scaffold which can be readily substituted with a variety of electrophiles using standard synthetic methods.2
In analogy to napthalenediimide3a-h and acridine-3- and 4-carboxamide3i-k heterocycles, we envisioned that the extended planar π surface of the indolo[3,2-b]carbazole chromophore could efficiently stack between the nucleobases of duplex DNA via an intercalative mode of binding,4a and thus derivatives of indolo[3,2-b]carbazole may function as high affinity nucleic acid ligands.4b Depending on the orientation of the planar chromophore in the intercalation complex, the two substituents on the nitrogen atoms could be positioned in either the same groove or in both the major and minor grooves of DNA, where non-covalent interactions may afford a level of sequence specificity. To test this hypothesis, we have synthesized a series of N-substituted derivatives of indolo[3,2-b]carbazole (Figure 1) and evaluated their affinity for duplex DNA via ultraviolet and fluorescence spectroscopies.
Figure 1.
Indolo[3,2-b]carbazole (1) and synthetic derivatives studied (2, 3).
It is well established that ligands containing positively charged residues have a high affinity for duplex nucleic acids due to favourable electrostatic interactions with the negatively charged phosphate backbone and electronegative atoms in the major and minor grooves developed upon initial association with the biopolymer.5 Thus the primary series of indolocarbazole derivatives targeted for synthesis (2a-e and 3a-e) contain basic nitrogen atoms.6
Both Waring7 and McGee8 have previously demonstrated that the presence of a carbohydrate moiety on an intercalating chromophore contributes positively and significantly to the binding association with DNA. The bis-C-aryl glycoside natural product altromycin B has been shown by NMR studies to associate with DNA via a helix-threading mode of binding, with carbohydrate moieties positioned in opposite grooves of the duplex,9 where non-covalent interactions between functional groups present on the sugar and residues in the major and minor grooves are established. Given the facile substitution of the opposing nitrogen atoms on the indolo[3,2-b]carbazole chromophore, we considered that mono- and bis-carbon glucoside derivatives (2f and 3f) could also easily be prepared, allowing a direct comparison of the DNA-binding affinity of ligands containing one or two attached glycosyl moieties.
The synthesis of the parent chromophore 1 was accomplished from 3,3’-diindolylmethane in 82% yield via an acid-promoted cyclization as described by Pindur and Mueller (Scheme 1).1 Alkylation of 1 under basic conditions (NaH, DMF) with 2-dimethylamino-1-chloroethane·HCl provided 3a (71%) when more than 2 equivalents of alkylating agent were used, and separable mixtures of 2a (16%) and 3a (58%) when less than 2 equivalents of alkylating agent were employed.2 In an analogous fashion, substrates 2d (22%), 3d (35%) and 2e (44%), 3e (23%) were prepared by reaction of 1 with N-methyl-2-chloromethyl imidazole hydrochloride and 2-bromomethyl-pyridine, respectively, in the presence of sodium hydride. Reaction of the monosodium salt of 1 in DMF with 2-chloro-ethylimidazole hydrochloride or 2-chloroethylpyridine-hydrochloride provided substrates 2b and 2c in 18% and 7% yields, respectively. Low yields were encountered in these alkylation reactions due to elimination side reactions, and as a result double alkylation of 1 under basic conditions could not be achieved with these reagents. However, 1•2Na reacted with excess TBS-protected 2-bromoethanol to provide bis-silyl ether 4 in 92% yield, which was then treated with TBAF to provide alcohol 3b in 98% yield.
Scheme 1.
Synthesis of indolo[3,2-b]carbazoles 2a-e and 3a-3e.
The synthesis of indolo[3,2-b]carbazole derivatives 2f and 3f required preparation of an appropriate carbohydrate alkylating agent (Scheme 2). Glucal acetonide 510 was deprotonated with Schlosser’s base in THF at −78°C for 1 hour11 and then DMF was added to furnish the corresponding aldehyde 6 in 92% yield. Reduction of the aldehyde to the allylic alcohol was accomplished in 88% yield by treatment with L-selectride in THF at −78°C. Mesylation of the derived alcohol (MsCl, THF, Et3N, 0°C) then provided a reactive alkylating agent which smoothly coupled with the sodium salt of 1 to provide 10 in 80% yield. Glycoside 10 was then subjected to hydroboration/oxidation, silyl ether deprotection, and acetonide hydrolysis to furnish the bis glycoside 3f in 59% overall yield.
Scheme 2.
Synthesis of indolo[3,2-b]carbazole derivatives 2f and 3f.
The synthesis of monoglycoside 2f required preparation of the mono-Boc indolo[3,2-b]carbazole 9, and this was accomplished in two steps (79%) according to the protocol of Bergman.2 Combination of the sodium salt of 9 with 8 furnished glycoside 11 in 37% yield. Hydroboration/oxidation, silyl ether deprotection with TBAF, and acetonide hydrolysis gave the penultimate Boc-protected glycoside, which upon carbamate thermolysis in water gave 2f in 60% overall yield from 11.
The binding of the indolo[3,2-b]carbazole derivatives to duplex DNA was explored by fluorescence and UV spectroscopies. Excitation of the most basic derivative 3a (0.064 μM in 10 mM Tris-EDTA, pH=5.48) at 320 nm in the presence of increasing concentrations of calf thymus (CT) DNA resulted in a corresponding decrease of the emission intensity at 396 nm (Figure 2), indicative of the quenching of the fluorescence of 3a upon binding to DNA.12 Monitoring the UV absorbance of 3a at 336 nm upon titration with CT-DNA in Tris-EDTA buffer revealed a distinct hypochromic shift along with a red shift to 342 nm (isosbestic point=340 nm), observations suggestive of an intercalative mode of binding (see below).
Figure 2.
Fluorescence spectra of 3a in the presence of varying concentrations of CT-DNA at pH 5.48, [3a] = 6.4 × 10−8 M, [CT-DNA] = 0.02, 0.03, 0.05, 0.08, 0.13, 0.20, 0.30, 0.40, 0.60, 0.90, 1.20, 2.4, 5.4, and 8.40 × 10−6 M, respectively.
To compare the strengths of binding of compounds 2a-e and 3a-e to CT DNA, the competitive ethidium displacement technique was employed to obtain apparent association constants (Kapp).13 From this analysis (Table 1) the tightest binding indolo[3,2-b]carbazole was bisimidazole derivative 3d with Kapp=1.23 × 107 M−1, a value which is more than three-fold higher than that obtained for monoimidazole derivative 2d (Kapp=3.49 × 106 M−1). Similarly, the apparent binding constant for bis(dimethylaminoethyl) derivative 3a (4.52 × 106 M−1) is approximately threefold greater than that obtained for the mono(dimethylaminoethyl) derivative 2a (1.47 × 106 M−1). Interestingly, compounds 2d and 2e have lower apparent binding constants for CT DNA than their homologs 2b and 2c, respectively, indicating the importance of the additional methylene unit between the indolocarbazole chromophore and the imidazole or pyridine moiety for tighter DNA binding. Finally, all derivatives lacking basic amine groups (3b, 2f, and 3f) display relatively poor binding to CT DNA, suggesting that positively charged groups on the indolocarbazole side chains significantly enhance the association with DNA.
Table 1.
Comparison of the binding of 2a-2f and 3a-3f to CT DNAa by ethidium displacementa and thermal denaturation techniques.b
Ligand | Kapp (×106M−1)a | Δ TM(°C)b |
---|---|---|
2a | 1.47 (±0.11) | 1 |
2b | 6.26 (±0.63) | – d |
2c | 2.06 (±0.14) | (−47)d,20 |
2d | 3.49 (±0.65) | (−50)d,20 |
2e | 0.70 (±0.45) | – d |
2f | <0.01 | – c |
3a | 4.52 (±0.51) | 6 |
3b | <0.01 | – c |
3d | 12.3 (±0.13) | (−18)d,20 |
3e | 3.77 (±0.39) | – d |
3f | <0.01 | 0 |
Kapp values obtained by the competitive ethidium displacement method (2mM NaOAc, 9.3 mM NaCl, 0.1mM Na2EDTA, pH=5.0), where Kapp=Ke × 1.26/C50 and Ke=2.1 × 106M−1. 13
TM values obtained (10 mM Tris-EDTA, 0.1M NaCl, pH=5.0) from first derivative analysis (ΔA/ΔTvs. ΔT) of the sigmoidal melting curves (A vs T); ΔTM = TM(ST DNA+ligand)− TM(ST DNA).14
Compounds 2f and 3b display poor solubility in the buffer used for the thermal denaturation studies.
Overlap of the UV absorbances of pyridine- and imidazole-substituted derivatives 2b-2e and 3d,e with DNA in the 260-280 nm region prevented accurate TM determinations.20
Thermal denaturation studies14 were also performed with compounds 2a-e and 3a-e and calf thymus (CT) DNA (10 mM Tris-EDTA, 0.1M NaCl, pH=5.0). Poor solubility in 0.1 M NaCl buffer hampered TM determinations for compounds 2f and 3b in the presence of CT DNA; furthermore, overlap of absorbances in the 260-280 nm region of the UV spectra of imidazole- and pyridine-substituted derivatives 2b-2d and 3d,e with CT DNA complicated Tm analysis for these compounds.20 Reliable data obtained for compounds 2a, 3a, and 3f confirmed the strong binding of 3a to CT DNA.
To investigate the importance of electrostatic interactions5 in the binding of the basic indolocarbazole derivatives to DNA, we evaluated the association of compound 3a with CT DNA under high- and low-salt conditions (Figure 3).4 Increasing NaCl concentrations from 0-0.01M results in a modest two-fold drop in 3a’s DNA binding affinity; however, at a salt concentration of 0.1M a >10 fold drop in affinity is observed. These data suggest that the positively charged protonated nitrogen atoms may be involved both in polar hydrogen bonding interactions in the grooves of DNA and in ionic interactions with the phosphate backbone.
Figure 3.
Binding isotherms for the titration of 3a with CT DNA in the presence of varying concentrations of NaCl. aBlack circle: [NaCl]=0.0 M, Ka=6.82(±0.86) × 106 M−1; bred square: [NaCl]=1.0 ×10−3M, Ka=6.69(±0.98) × 106 M−1; cgreen diamond: [NaCl]=1.0 ×10−2M, Ka=3.02(±0.37) × 106 M−1; dblue triangle: [NaCl]=1.0 ×10−1M, Ka=4.69(±0.36) × 105 M−1.
To further probe the DNA binding mode of the indolocarbazole derivatives, we assayed for ligand intercalation using the technique of viscometry.15 Increasing concentrations of 3a (Figure 4) in the presence of CT-DNA resulted in an increase in solution viscosity comparable to that obtained with reference compound ethidium bromide (EB), indicative of the helix lengthening and rigidification that occurs upon ligand intercalation.
Figure 4.
Effect of increasing amount of EB (●) and 3a (■) on the relative viscosity of CT-DNA. EB (●): [CT-DNA]= 4.81 × 10−4 M, [EB] =0.44, 2.61, 6.96, 11.3, 15.7, and 20.0 ×10−5M; 3a (■): [CT-DNA] = 2.00 × 10−4 M, [3a] = 0.78, 2.72, 6.60, 10.5, 14.4, and 18.2 × 10−5 M.
Depending on the orientation of the chromophore upon intercalation, the nitrogen substituents may be positioned in either the minor or major grooves or both. To check for major groove occupancy, a fixed concentration of 3a and CT DNA was titrated with increasing concentrations of the non-intercalating major-groove binding dye methyl green,16 and displacement of 3a from DNA was observed, as evidenced by fluorescence enhancement at 410 nm (Figure 5). In contrast, increasing concentrations of methyl green in the presence of 3a but in the absence of CT DNA showed no significant enhancement of fluorescence. The UV absorption spectra of 3a in the presence of polydA•2polydT or polydA•polydT show similar shifts in the 335 nm (λmax) region as compared to compound 3a in the absence of DNA, indicative of binding to both duplex and triplex forms. Since the third strand of polydA•2polydT occupies the major groove, this result implies that 3a can also associate with the minor groove of DNA. Indeed, titration of 3a and CT DNA with increasing concentrations of the minor groove-binding drug Hoechst 33342 gave rise to fluorescence enhancement, which is suggestive of minor groove contacts by the ligand.
Figure 5.
[a] Fluorescence spectra of 3a (0.12 ×10−6M) in the presence of CT DNA (0.37 × 10−6M) and varying concentrations of methyl green: 0.00, 0.76, 1.52, 3.04 ×10−6 M. [b] Fluorescence spectra of 3a (0.12 × 10−6 M; no CT DNA) in the presence of varying concentrations of methyl green: 0.00, 0.76, 1.52, 3.04 ×10−6 M.
We examined the sequence selectivity of 3a by measuring its binding to a series of sequence-diverse DNA hairpins 5’-CGXXXXXC-AAAAA-GXXXXXCG-3’ (Table 2).17 A plot of the change in fluorescence versus molar equivalents of ligand for a given hairpin provides a titration curve from which the stoichiometry of binding may be derived; performing this analysis for 3a in the presence of hairpins 5’-AATCT-3’ (5’-TTAGA-3’) and 5’-TTTTT-3’ (5’-AAAAA-3’) gave stoichiometries of 1.10 and 0.85, respectively, indicating ~1:1 binding. Titration of a fixed concentration of hairpin with 3a yields a sigmoidal binding curve, fitting analysis of which provides Ka. As can be seen in Table 2, tightest binding was observed to the hairpins containing the mixed sequences 5’-AGAGA-3’, 5’-GGTAG-3’, 5’-AACGG-3’ and 5’-TATAT-3’, with Ka values two- to three-fold higher than those for the hairpins containing continuous AT and GC tracts. With the exception of 5’-AGAGA-3’, all of the tightest bound hairpins contain purine-pyrimidine steps. Purine-pyrimidine steps are more weakly π−stacked than purine-purine/pyrimidine-pyrimidine steps and thus intercalators typically display a preference for binding these sequences.19,22
Table 2.
Binding of 3a to DNA hairpins:
hairpin | TGTGT ACACA |
TATAT ATATA |
AATCT TTAGA |
GGGGG CCCCC |
AAAAA TTTTT |
| |||||
Ka
(×107M−1) |
29.2 (±2.6) |
36.5 (±5.5) |
23.5 (±3.2) |
15.7 (±1.6) |
17.3 (±1.5) |
| |||||
hairpin | GCGCG CGCGC |
AACGG TTGCC |
GGTAG CCATC |
AGCCT TCGGA |
AGAGA TCTCT |
| |||||
Ka
(×106M−1) |
23.9 (±3.2) |
41.3 (±4.7) |
47.6 (±3.9) |
21.8 (±1.8) |
48.1 (±5.2) |
|
Finally, a preliminary assessment of the cell permeability and cytotoxicity of the indolo[3,2-b]carbazoles was performed with compound 3a. Treatment of liver carcinoma (Hep2G) cells with 3a (at 1.38 and 55.2 μM concentrations) resulted in significant cellular uptake within 4.5 hours, as evidenced by residual fluorescence (at 410 nm) of the medium after extensive buffer washing (including heparin, low pH, high salt, and trypsin treatment) and cell lysis (see Supporting Information) as compared to controls lacking 3a. Furthermore, the viability of acute leukemia monocytes in the presence and absence of 3a was assayed using the trypan blue exclusion test.18 It was found that within 20 hours <10% of cells treated with 3a (1.3×10−5M) were still viable. These early results suggest a generalized cytotoxicity of cell-permeable indolo[3,2-b]carbazole deriv-atives likely associated with tight DNA binding. Detailed cytotoxicity studies (including MTT/MTS assays)21 are currently in progress and will be reported in due course.
Conclusions
A variety of N-monosubstituted and N,N-disubstituted indolo[3,2-b]carbazoles have been prepared and evaluated for their binding to duplex DNA. It has been found that derivatives possessing one or two basic N-substituted groups bind tightly to DNA, with bis-imidazole compound 3d displaying the highest apparent affinity. Evidence for an intercalative mode of DNA binding has been established, and it appears likely that the N-alkyl substituents project into the minor and/or major grooves of the double helix. Indolo[3,2-b]carbazole 3d prefers to bind sequences of mixed composition and those containing purine-pyrimidine steps, perhaps due to the greater ease of intercalation at these sites. The cell permeability and cytotoxicity of these compounds have been preliminarily investigated, and the results may be of interest for the materials industry seeking to utilize such substituted photoactive chromophores in the design of organic electronics.
Supplementary Material
Acknowledgements
We thank the National Institutes of Health (SC3 GM 096899-01) for their generous support of our research program. We also thank Dr. Jheem Medh (CSUN Department of Chemistry and Biochemistry) and Dr. Rheem Medh (CSUN Department of Biology) for their assistance with the cell permeability and viability studies.
Footnotes
Electronic Supplementary Information (ESI) available: [details of any supplementary information available should be included here]. See DOI: 10.1039/c000000x/
References
- 1.Pindur U, Mueller J. Arch. Pharm. 1987;320:280. [Google Scholar]
- 2.Yudina LN, Bergman J. Tetrahedron. 2003;59:1265. [Google Scholar]
- 3.a McKnight RE, Reisenauer E, Pintado MV, Polasani SR, Dixon DW. Bioorg. Med. Chem. Lett. 2011;21:4288. doi: 10.1016/j.bmcl.2011.05.069. Naphthalenediimides: [DOI] [PubMed] [Google Scholar]; b Hampel SM, Sidibe A, Gunaratnam M, Riou JF, Neidle S. Bioorg. Med. Chem. Lett. 2010;20:6459. doi: 10.1016/j.bmcl.2010.09.066. [DOI] [PubMed] [Google Scholar]; c Lee J, Guelev V, Sorey S, Hoffman DW, Iverson BL. J. Am Chem Soc. 2004;12:14036. doi: 10.1021/ja046335o. [DOI] [PubMed] [Google Scholar]; d Gianolio DA, Segismundo JM, McLaughlin LW. Nucleic Acids Res. 2000;28:2128. doi: 10.1093/nar/28.10.2128. [DOI] [PMC free article] [PubMed] [Google Scholar]; e McKnight RE, Gleason AB, Keyes JA, Sahabi S. Bioorg. Med. Chem. Lett. 2007;17:1013. doi: 10.1016/j.bmcl.2006.11.038. [DOI] [PubMed] [Google Scholar]; f Ohtsuka K, Komizo K, Takenaka S. J. Organomet. Chem. 2010;695:1281. [Google Scholar]; g Mazzitelli CL, Chu Y, Reczek JJ, Iverson BL, Brodbelt JS. J. Am. Soc. Mass. Spectrom. 2007;18:311. doi: 10.1016/j.jasms.2006.09.021. [DOI] [PMC free article] [PubMed] [Google Scholar]; h Guelev V, Lee J, Ward J, Sorey S, Hoffman DW, Iverson BL. Chem. Biol. 2001;8:415. doi: 10.1016/s1074-5521(01)00013-8. [DOI] [PubMed] [Google Scholar]; i Howell LA, Gulam R, Mueller A, O’Connell MA, Searcey M. Bioorg. Med. Chem. Lett. 2010;20:6596. doi: 10.1016/j.bmcl.2010.09.128. Acridine-3- and 4-carboxamides: [DOI] [PubMed] [Google Scholar]; j Gooch BD, Beal PA. J. Am. Chem. Soc. 2004;126:10603. doi: 10.1021/ja047818v. [DOI] [PubMed] [Google Scholar]; k Bourdouxhe-Housiaux C, Colson P, Houssier C, Waring MJ, Bailly C. Biochemistry. 1996;35:4251. doi: 10.1021/bi9528098. [DOI] [PubMed] [Google Scholar]
- 4.a Sajewicz W, Dlugosz A. J. Appl. Toxicol. 2000;20:305. doi: 10.1002/1099-1263(200007/08)20:4<305::aid-jat659>3.0.co;2-p. [DOI] [PubMed] [Google Scholar]; b Jia T, Xiang J, Wang J, Guo P, Yu J. Org. Biomol. Chem. 2013;11:5512. doi: 10.1039/c3ob40799c. [DOI] [PubMed] [Google Scholar]
- 5.Misra VK, Honig B. Proc. Natl. Acad. Sci. 1995;92:4691. doi: 10.1073/pnas.92.10.4691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Anslyn EV, Dougherty DA. Modern Physical Organic Chemistry. University Science Books; Sausalito: 2006. p. 281. Approximate pKb values for pyridine (8.77), imidazole (6.90) and trimethylamine (4.20) are derived from the corresponding conjugate acid pKa values in water: [Google Scholar]
- 7.Bailly C, Colson P, Houssier C, Rodrigues-Periera E, Prudhomme M, Waring MJ. Mol. Pharmacol. 1998;53:77. doi: 10.1124/mol.53.1.77. C. [DOI] [PubMed] [Google Scholar]
- 8.Tse-Dinh YC, McGee LR. Biochem. Biophys. Res. Commun. 1987;143:808. doi: 10.1016/0006-291x(87)90320-2. [DOI] [PubMed] [Google Scholar]
- 9.Hansen MR, Hurley L. J. Am. Chem. Soc. 1995;117:2421. [Google Scholar]
- 10.Dotz KH, Otto F, Nieger M. J. Organomet. Chem. 2001;621:77. [Google Scholar]
- 11.a Paquette LA, Oplinger JA. Tetrahedron. 1989;45:107. [Google Scholar]; b Friesen RW, Sturino CF. Science of Synthesis. 2006;8a:841. [Google Scholar]
- 12.McGhee J, von Hippel P. J. Mol. Biol. 1974;86:469. doi: 10.1016/0022-2836(74)90031-x. Analysis of the fluorescence data by least squares fitting analysis gave Ka= 3.8(±0.4) × 106 M−1. For non-linear regression techniques, see. [DOI] [PubMed] [Google Scholar]
- 13.Morgan AR, Lee JS, Pulleyblank DE, Murray NL, Evans DH. Nucleic Acids Res. 1979;7:547. doi: 10.1093/nar/7.3.547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wilson WD, Tanious FA, Fernandez-Saiz M, Rigl CT. Methods in Mol. Biol. Drug-DNA Interact. Protoc. 1997;90:219. doi: 10.1385/0-89603-447-X:219. [DOI] [PubMed] [Google Scholar]
- 15.Dedon PC. Curr. Protoc. Nucleic Acid. Chem. John Wiley and Sons; New York: 2000. 8.1.1. [Google Scholar]
- 16.Kim SK, Norden B. FEBS Letters. 1993;315:61. doi: 10.1016/0014-5793(93)81133-k. [DOI] [PubMed] [Google Scholar]
- 17.a Tse WC, Boger DL. Acc. Chem. Res. 2004;37:61. doi: 10.1021/ar030113y. [DOI] [PubMed] [Google Scholar]; b Boger DL, Tse WC. Bioorg. Med. Chem. 2001;9:2511. doi: 10.1016/s0968-0896(01)00243-7. [DOI] [PubMed] [Google Scholar]
- 18.Strober W. Curr. Protoc. Immunol. 2001. Appendix 3, Appendix 3b.
- 19.a Tse WC, Boger DL. Chem. Biol. 2004;11:1607. doi: 10.1016/j.chembiol.2003.08.012. [DOI] [PubMed] [Google Scholar]; b Johnson DS, Boger DL. DNA binding agents. In: Murakami Y, editor. Comprehensive Supramolecular Chemistry. Elsevier; Oxford: 1996. p. 73. [Google Scholar]
- 20. The apparent TM’s of CT DNA in the presence of compounds 2c, 2d, and 3d were lower than in their absence, suggestive of preferential binding to single-stranded DNA. See reference 14.
- 21.a Mossmann T. J. Immunol. Methods. 1983;65:55. doi: 10.1016/0022-1759(83)90303-4. [DOI] [PubMed] [Google Scholar]; b Cory AH, Owen TC, Barltrop JA, Cory JG. Cancer Commun. 1991;3:207. doi: 10.3727/095535491820873191. [DOI] [PubMed] [Google Scholar]
- 22. Hairpin melting data in the presence and absence of 3a roughly corroborate the trends observed in Table 2 (see Supporting Information)
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.