Alkali Cation- and Borate-Binding States of Tartrolon Antibiotics

Abstract

The borate-binding polyketide macrolide natural products are a long-known family of bacterial antibiotics and antiparasitic agents. Among these, tartrolon E is highly potent and selective in killing eukaryotic parasites while sparing mammalian cells. However, it has been challenging to obtain, fully chemically define, and formulate the tartrolons. Here, we describe a streamlined route to obtain pure tartrolon E as a highly crystalline material. The method yielded crystals of tartrolon E, the analysis of which revealed the stable chelation of a sodium counterion. Using this chemically defined material, additional experiments permitted quantitative cation exchange with alkali-metal cations, suggesting a relative binding affinity of Li⁺ > Na⁺ > K⁺. In cases where complex mixtures of boronated/deboronated tartrolons are obtained, we developed methods to cleanly deboronate tartrolon E, yielding tartrolon D, and to reintroduce the boron atom back to the complex. Overall, we demonstrate practical methods to deliver chemically defined complexes of tartrolon E, which will facilitate further study of the intriguing biological activities of this potently bioactive macrolide family and enable the preclinical development of these important antiparasitic and antibiotic agents.

Boron is an essential trace nutrient for vascular plants − and may be essential for other organisms. − Medicinal chemists have harnessed the unique properties of boron to develop several FDA-approved drugs that target cancer, infectious disease, and entopic dermatitis. Microorganisms also incorporate boron into small molecules with wide-ranging biological functions, including quorum-sensing signals, pigments, and pharmacologically active secondary metabolites. Several natural products feature the Böeseken complex, in which borate is held in a tetrahedral coordination between four alcohols or α-hydroxy acids at the center of an organic macrolide. Borate macrolides are produced by diverse bacteria living in disparate environments.

Boromycin (1) was the first boron-containing biomolecule to be characterized. Originally isolated in 1967 from the African soil-derived microbe Streptomyces antibioticus, boromycin and its desvalinoboromycin analogue have been reisolated from other Streptomyces strains through bioassay-guided discovery efforts. Boromycin displays broad biological activity against Gram-positive pathogens, , Toxoplasma and Cryptosporidium parasites, and HIV replication. Subsequently, several other boron macrolides have been described. Aplasmomycin (2) was isolated from Streptomyces griseus collected in shallow sea mud and exhibited activity against Gram-positive bacteria in vitro and against Plasmodium in vivo. In contrast to boromycin, the aplasmomycin macrocycle is symmetrical, consisting of two identical PKS chains. Mono- and diacetylated aplasmomycins were isolated from Streptomyces griseus NCIB 11371. Borophycin (3), which is also symmetrical, was isolated as a potent cytotoxin from the cyanobacterium Nostoc linckia. Notably, this compound was characterized as a mixture of two forms, including one in which methanol replaces one of the hydroxyl oxygens in coordinating a sodium counterion (Figure ).

1.

Structures of Böeseken complex-containing borate macrolides from bacterial sources.

Tartrolons A (4) and B (5) were first isolated from the myxobacterium Sorangium cellulosum. Tartrolon A, lacking boron, was difficult to isolate, requiring the use of stainless-steel tanks (compared to borate-leaching glass-lined vessels) during the fermentation process. Further, tartrolon A was isolated as a mixture of stereoisomers (A₁–A₃). The boronated tartrolon B crystal structure showed borate and sodium coordination nearly identical to that of borophycin, mirrored by the conserved structure and absolute configuration of the coordination centers in the two molecules. Interestingly, it was reported that attempts to form the cesium salt were unsuccessful, as the crystal obtained always contained a sodium ion. Additionally, adding excess potassium to Listeria monocytogenes assays significantly decreased tartrolon B activity, implicating tartrolon B as a potential K⁺ ionophore. Tartrolon C (6), differing from tartrolon B by addition of hydroxyl groups on C4 and C4′, was isolated based upon insecticidal activity from a soil-derived Streptomyces strain, inhibiting the hypoxia inducible factor-1 (HIF-1) pathway.

Of the borate macrolides isolated to date, tartrolons D (7) and E (8) have arguably shown the greatest promise for development. Tartrolon D was originally isolated from marine actinomycete Streptomyces sp. MDG-04-17-069. Only minor traces of the boronated version were observed and believed to be artifacts from storage in glass vials. The boronated version, tartrolon E, was isolated and characterized from the marine shipworm symbiont gammaproteobacterium Teredinibacter turnerae T7901, which produced a mixture of tartrolons D and E. Tartrolon E has since been the focus of numerous studies investigating its potent activity against parasites. The compound has potent, broad-spectrum antiapicomplexan activity, targets asexual and sexual stages of Cryptosporidium parvum and Plasmodium falciparum, , and blocks host cell invasion by Toxoplasma gondii while having low toxicity to mammalian cell lines.

The long history of literature on borate-containing macrolides demonstrates a few of the difficulties associated with the compounds. Several are regularly isolated in a mixture of forms based on the presence or absence of borate, and further racemization or degradation can occur when borate is absent. Although complexes with other alkali-metal cations (Cs⁺ and Rb⁺) and Ag⁺ have been prepared to aid crystallographic characterization of these complexes, little has been reported on the dynamics of borate and counterion removal and replacement. Here, we optimized the growth, extraction, and purification of tartrolon E from Teredinibacter turnerae to obtain highly pure material for subsequent investigation of the removal and replacement of borate and a series of counterions to better understand the dynamics of this structural class.

Results and Discussion

Improved Method for Tartrolon Purification

The preparative scale production, purification, and characterization of boron-containing natural products pose several challenges. The complexes are often isolated as mixtures of borate-intact and hydrolyzed products, which makes HPLC purification difficult. Furthermore, mildly acidic or strongly nucleophilic conditions can result in further hydrolysis and degradation of the products, which are difficult to reverse and result in diminished yields for downstream biological testing. Obtained tartrolon forms are often difficult to dissolve, complicating dissolution and formulation. Finally, copurification of tartrolon E and other boron-containing natural products with fatty acids necessitates multiple rounds of purification, further hampering isolation and greatly reducing isolated yields.

To solve these problems and increase the isolated yield of a pure, single form of tartrolon E, a streamlined fermentation, extraction, and purification method was developed. To ensure adequate borate was available, boric acid was supplemented in the fermentation broth to a final concentration of approximately 150 μM. This included borate from artificial seawater, the minerals and metals mix used to prepare media, and supplemented boric acid. During T. turnerae growth, media pH decreases over time, especially when cellulose is used as the carbon source. Additionally, in our experience, later stage T. turnerae cultures begin to produce slime and are no longer able to be pelleted by centrifugation. Cultures were thus harvested by centrifugation after 3–4 days, before the pH dropped below 7.0 or slime production occurred (Figure S1). The cell pellet was frozen and lyophilized and then exhaustively extracted with 2:1 CH₂Cl₂/MeOH. The extract was dried in vacuo and partitioned between ddH₂O and CH₂Cl₂. The CH₂Cl₂ partitions were combined and washed with saturated NaCl, providing excess Na⁺ as a stabilizing cation. The organic layer was then dry loaded onto C₁₈ resin and purified by reverse phase flash chromatography on a C₁₈ column that had never been used under acidic conditions, yielding a single peak for tartrolon E by UV–vis detection (Figure S2).

The collected tartrolon E fractions still contained a large amount of fatty acids, which complicates downstream biological testing, solubility, and formulation. To remove these, the obtained material was triturated in cold hexanes and filtered over a fine glass frit filter under a vacuum. The solid material was washed with hexanes three times, resulting in pure tartrolon E (Figure S3). This procedure resulted in an isolated yield of 23.4 mg/L of culture, while only necessitating a single chromatographic step, improving isolated yield by at least 3-fold and greatly improving process efficiency. Importantly, this material contains only a single form of tartrolon E with a sodium counterion and is readily dissolved in a range of organic solvents, improving the potential for formulation. By contrast, in previous work in our hands, processed tartrolons were much more difficult to dissolve and formulate. The final product was confirmed by NMR and X-ray crystallography to be a single chemical entity with intact boron and a sodium counterion.

Crystal Structure of Tartrolon E

Isolated tartrolon E from the improved purification method was dissolved in 9:1 CH₂Cl₂/MeOH, and a few drops of ddH₂O were added. Colorless plank crystals were obtained after 24 h of evaporation, and the 3D structure of tartrolon E was unambiguously confirmed based on the anomalous dispersion of Cu K_α radiation (deposited as CCDC 2473545) (Figure ).

2.

Crystal structure of tartrolon E with sodium counterion (CCDC 2473545). Hydrogens are not shown for the sake of clarity. (A) “Top” view down the Na–B axis with sodium (yellow sphere) on top. (B) Side view showing tetrahdedral coordination of boron (pink) at the bottom of the disk. (C) “Bottom” view up the B–Na axis with boron on top. (D) Side view.

The overall structure is highly symmetrical and bears a striking resemblance to several other boron-containing polyketides previously described. As expected, the boron is coordinated by four oxygens O3, O4, O3′, and O4′ in the middle of the disk. The sodium counterion is coordinated through six bonds from O2, O3, O7, and O7′. When compared to tartrolon C, the most closely related compound, the structures are almost identical except for the absence of the hydroxyl group on C4 and methyl group on C8 (Figure S8).

Borate Exchange

To address the problematic equilibration that has been observed between tartrolons D and E, and to ease spectral characterization of complex tartrolon mixtures, we developed methods to both abstract and incorporate the boron atom.

Boric acid complexes demonstrate complex equilibrium behavior with rapid exchange of boron. Given that the tetrahedral Böesekin complexes formed between boric acid and organic polyols typically have pK _as in the 8.5–9.2 range, protonation of the borate rapidly catalyzes borate exchange in solution. This is true for tartrolon E, wherein the borate is extremely labile under acidic conditions. Sequential addition of 0.5 to 10 equiv of trifluoroacetic acid (TFA) in methanol led to rapid loss of the borate (Figure ). After the addition of 10 equiv of TFA, an equilibrium ratio of 2.9:1 tartrolon D:E is established.

3.

Elimination of borate in the presence of an acid. (A) ¹H NMR spectra of tartrolon E (500 MHz, CD₃OD) followed by the addition of 0.5–10 equiv of trifluoroacetic acid. (B) Corresponding ¹¹B-NMR spectra demonstrating the deboronation of tartrolon E and formation of trimethylborate.

When complex mixtures of boronated/deboronated tartrolons are isolated, it may be convenient to fully deboronate and/or reintroduce boron. While TFA is capable of catalyzing deboronation, the reaction may not proceed to completion, and we have noted significant compound decomposition over longer time periods. KHF₂ with a catalytic amount of TFA is capable of quantitatively deboronating tartrolon E, delivering tartrolon D (Figure B). To reboronate tartrolon D, we attempted several experimental protocols with trimethyl borate as previously described in the synthesis of related boron macrodiolide natural products. − Under these conditions, only partial reboronation was observed even with a large excess of trimethyl borate. Treatment with triisopropyl borate (B­(iOPr)₃), however, reliably gave complete reboronation to revert tartrolon D to E (Figure C). The ¹H chemical shifts of the reboronated tartrolon E are identical to those of the isolated natural product, indicating that the counterion present is Na+.

4.

Complete deboronation and reboronation of tartrolon E in CD₃OD. (A) ¹H NMR spectra of isolated natural product tartrolon E (top), quantitative deboronation with KHF₂ and cat. TFA (middle), and reintroduction of boron to reform tartrolon E using triisopropyl borate (B­(iOPr)₃) (bottom). (B) Corresponding ¹¹B-NMR spectra for respective conditions in panel A.

Cation Exchange

Tartrolon E has been isolated as only the sodium salt. Often, boron macrolide natural products are isolated and characterized but without definitive characterization of which cation accompanies the borate anion. Given previous reports indicating that tartrolons may act as potassium ionophores, a general method to characterize the cation binding affinity of these complexes was developed.

Without an appreciation of the relative binding affinity to the cations, we first sought to exchange the cation present in the natural product, which was Na⁺ in the crystal structure. Leveraging the insolubility of NaBr in acetone, LiBr was added to a solution of tartrolon E in acetone-d ₆ (Figure ), which quantitatively converted the material to the Li⁺ salt (Figure ). Likewise, the addition of an excess (10 equiv) of the more soluble KI to the sodium salt quantitatively converted tartrolon E from the sodium to potassium form. This protocol also allowed us to qualitatively examine the relative stability of the alkali metal salts (Table ). For example, the addition of excess NaI or KI to the Li⁺ salt demonstrated no cation exchange, suggesting that the lithium complex is more stable than the sodium complex. Alternatively, addition of excess NaI to the K⁺ salt rapidly exchanged Na⁺ for K⁺, suggesting the relative stability order is Li⁺ > Na⁺ > K⁺. During salt exchange, differences in ¹H NMR data reflected the size of the bound cation, with Na⁺ complex shifts intermediate between the K⁺ and Li⁺ values. Several of the largest ¹H NMR shift changes were observed not immediately adjacent to the cation, but instead at key structural positions such as H₂₀ and H₈, dihedral angles of which might be expected to change as smaller or larger alkali ions are accommodated in the tartrolon cation binding pocket.

5.

Counterion exchange in tartrolon E in acetone-d ₆. ¹H NMR of Na⁺ salt of tartrolon E isolated from T. turnerae culture (top), ¹H NMR of K⁺ salt of tartrolon E resulting from the addition of excess KI (middle), and ¹H NMR of Li⁺ salt of tartrolon E resulting from the addition of LiBr (bottom).

1. Significant ¹H NMR Changes in Tartrolon E Salts in ppm (Δ from Li Salt), in Acetone-d ₆ .

	lithium	sodium	potassium
H7	4.11	3.98 (−0.13)	3.88 (−0.23)
H8a	2.3	2.42 (+0.1)	2.51 (+0.2)
H10b	2.72	2.68 (−0.04)	2.60 (−0.12)
H20	4.78	4.67 (−0.11)	4.62 (−0.16)

Conclusion

This study sought to optimize the purification of the boron-containing natural product tartrolon E and use the resulting material to investigate the boron and cation binding characteristics of Böeseken complex macrolides. The streamlined purification process eliminated the coproduction of tartrolon D and enabled the lab-scale isolation of over 1 g of a single, stable form of tartrolon E. The material is spectroscopically pure and formed single crystals for X-ray analysis. Unsurprisingly, the resulting 3D structure is almost identical with that of tartrolon C and other symmetrical boron-containing macrolides.

Due to the reported difficulties with characterizing the dynamics of borate incorporation, as well as counterion coordination, methods were developed to quantitatively remove and replace boron and exchange the cation. Based upon the degree of challenge seen in the literatureand in our handsin controlling the yield and purity of boronated macrolide antibiotics, we believe that this was the limiting problem previously preventing this compound class from being developed as drug candidates. To circumvent this challenge, we developed a model for boronate macrolide chemistry that greatly improves control over producing highly pure, well-defined complexes in high yield (Scheme ). The ease of preparation and formulation of the resulting materials will help with biological assays and preclinical development of tartrolon and other boronate antibiotics moving forward. This may be especially important in treating the parasitic infections that are the major targets of tartrolon E.

1. Summary of Tartrolon Ion Exchange Chemistry.

In addition, the cation binding preferences of various boron macrolides are worth investigating and optimizing moving forward. Two applications for lithium binding ionophores are particularly promising at the current moment. In the first, because of the importance of lithium ion battery technology, it has become increasingly important to enrich the ion from more dilute sources such as seawater. In the second, lithium limitation has recently been implicated in Alzheimer’s disease. Thus, further exploration of lithium-binding agents such as tartrolons and others is a priority.

Experimental Section

General Experimental Procedures

NMR spectra were collected using a Varian 500 MHz NMR spectrometer with a 5 mm Varian HCN Oneprobe for proton detected experiments and an Agilent DirectDrive 500 MHz instrument for boron detected experiments. Residual signals from solvents were used for referencing. Flash chromatography was performed using a Teledyne NextGen 300 apparatus (Teledyne, CA, USA). X-ray analysis was performed on an XtaLAB Synergy R, DW system, and HyPix diffractometer.

Fermentation and Isolation of Natural Tartrolon E (1)

Shipworm basal medium (SBM) was formulated as previously described. Glycerol stocks of Teredinibacter turnerae T7901 were revived on SBM agar plates supplemented with 0.2% cellulose powder and incubated at 30 °C. After individual colonies formed, single colonies were excised from the plate and used to inoculate seed cultures (50 mL of SBM in 125 mL Erlenmeyer flasks). Cultures were incubated at 30 °C with shaking at 170 rpm for 1 day. Fernbach flasks (2.8 L) containing SBM cellulose (1 L) were supplemented with boric acid to f/c = 150 mM, and 1 mL of seed culture was used to inoculate cultures. Cultures were grown at 30 °C for 4 days with shaking at 100 rpm until a dense, pale yellow “growth ring” formed around the edge of the flask.

Cultures were harvested by centrifugation (7068 × g, 4 °C, 25 min), and cell pellets were frozen at −80 °C and lyophilized. Lyophilized pellets were extracted with 2:1 CH₂Cl₂:methanol (3 × 400 mL for 6 L of growth) and then dried in vacuo on a rotary evaporator. The extract was redissolved and partitioned between equal volumes of H₂O and CH₂Cl₂. The CH₂Cl₂ partitions were combined and washed with a saturated NaCl. Loose C₁₈ resin was added as a slurry in methanol (approximately 500 μL per 1 L of bacterial growth) to the washed CH₂Cl₂ fractions and dried to onto resin. The loaded resin was used to introduce the sample to a C₁₈ flash column (40 g, 35 mL/min) and eluted with a linear gradient from 40% acetonitrile in water to 100% ACN over 15 min followed by an isocratic hold at 100% acetonitrile. Fractions containing tartrolon were pooled and dried in vacuo. The resulting fraction was triturated in n-hexanes and vacuum filtered over a glass frit filter. The powder was washed three times with excess hexane to yield pure tartrolon E.

Tartrolon E Li⁺: (¹H NMR, 500 MHz, acetone-d ₆): δ 6.13–6.03 (m, 4H), 5.87 (tt, J = 9.7, 3.9 Hz, 2H), 5.27 (dddd, J = 10.0, 7.1, 5.4, 1.9 Hz, 2H), 4.83–4.74 (m, 2H), 4.46 (s, 2H), 4.40 (tdd, J = 11.6, 4.8, 2.2 Hz, 2H), 4.19 (t, J = 2.0 Hz, 2H), 4.15–4.09 (m, 2H), 3.32 (dd, J = 18.5, 10.3 Hz, 2H), 2.72 (dd, J = 13.1, 4.8 Hz, 2H), 2.44 (q, J = 11.2 Hz, 2H), 2.37–2.27 (m, 4H), 1.99–1.80 (m, 8H), 1.75 (tdd, J = 13.2, 9.4, 3.7 Hz, 2H), 1.69–1.59 (m, 4H), 1.58–1.51 (m, 2H), 1.43–1.34 (m, 2H), 1.35–1.25 (m, 4H), 1.21 (d, J = 6.2 Hz, 6H), 1.20–1.15 (m, 1H), 0.98 (d, J = 6.5 Hz, 6H).

Tartrolon E Na⁺: (¹H NMR, 500 MHz, acetone-d ₆): δ 6.12–6.03 (m, 4H), 5.88 (dt, J = 13.8, 5.0 Hz, 2H), 5.28 (td, J = 10.9, 10.2, 5.7 Hz, 2H), 4.67 (dtd, J = 11.6, 6.8, 5.4 Hz, 2H), 4.47 (s, 2H), 4.42 (tdd, J = 11.4, 4.4, 2.1 Hz, 2H), 4.02–3.95 (m, 2H), 3.59 (t, J = 2.0 Hz, 1H), 3.34 (ddd, J = 18.8, 10.5, 1.4 Hz, 2H), 2.68 (dd, J = 13.3, 4.4 Hz, 2H), 2.51–2.39 (m, 4H), 2.39–2.27 (m, 2H), 2.08 (d, J = 1.5 Hz, 1H), 2.02–1.80 (m, 8H), 1.75–1.61 (m, 6H), 1.58–1.50 (m, 2H), 1.42–1.35 (m, 2H), 1.34–1.26 (m, 4H), 1.22 (d, J = 6.2 Hz, 6H), 1.20–1.17 (m, 1H), 0.99 (d, J = 6.5 Hz, 6H).

Tartrolon E K⁺: (¹H NMR, 500 MHz, acetone-d ₆): δ 6.10–6.00 (m, 4H), 5.85 (dt, J = 13.9, 5.0 Hz, 2H), 5.24 (td, J = 11.0, 10.2, 5.5 Hz, 2H), 4.62 (dq, J = 12.1, 6.2 Hz, 2H), 4.48 (s, 2H), 4.43 (tt, J = 11.7, 3.2 Hz, 2H), 3.94–3.84 (m, 2H), 3.79 (t, J = 2.0 Hz, 1H), 3.33 (dd, J = 18.8, 11.2 Hz, 2H), 2.60 (dd, J = 13.8, 3.7 Hz, 2H), 2.54–2.43 (m, 4H), 2.42–2.33 (m, 2H), 2.07 (d, J = 2.0 Hz, 1H), 2.01–1.77 (m, 8H), 1.73–1.58 (m, 6H), 1.58–1.50 (m, 2H), 1.39 (t, J = 13.4 Hz, 2H), 1.36–1.27 (m, 4H), 1.23–1.19 (m, 1H), 1.18 (d, J = 6.1 Hz, 6H), 0.99 (d, J = 6.5 Hz, 6H).

X-ray Crystallographic Data for Tartrolon E (1)

Single colorless plank-shaped crystals of tartrolon E-Na were obtained by recrystallization from CH₂Cl₂:MeOH:H₂O. A suitable crystal 0.23 × 0.07 × 0.04 mm³ was selected and mounted on a suitable support on an XtaLAB Synergy R, DW system, HyPix diffractometer. The crystal was kept at a steady T = 101(1) K during data collection. The structure was solved with the ShelXT structure solution program using the Intrinsic Phasing solution method and by using Olex2 as the graphical interface. The model was refined with version 2018/3 of ShelXL 2018/3 using Least Squares minimization.

Borate Removal Method 1

Varying amounts of TFA (0.5, 2, and 10 equiv) were added to a solution of tartrolon E (2 mg, 0.0024 mmol) in CD₃OD (0.5 mL), and the mixture was stirred for 10 min.

Borate Removal Method 2

Solutions of KHF₂ in D₂O (10 μL, 4.5M) and TFA in CD₃OD (1.2 μL, 0.1 M) were added to a solution of tartrolon E (2 mg, 0.0024 mmol) in CD₃OD (0.5 mL) and stirred for 10 min.

Boron Incorporation

After either Method 1 or 2 was followed to remove the boron, the solvent was removed under reduced pressure, and H₂O (100 μL), saturated NaHCO₃ (50 μL), and CHCl₃ (100 μL) were added. The layers were separated, and the aqueous layer was extracted with CHCl₃ (2 × 100 μL). The combined organic layers were dried with Na₂SO₄ and filtered, at which point B­(OiPr)₃ (5.5 μL, 0.024 mmol) was added. After the mixture was stirred for 10 min, the solvent was evaporated under reduced pressure.

Cation Exchange

To a solution of tartrolon E (2 mg, 0.0024 mmol) in acetone-d ₆ (0.5 mL) was added the appropriate alkali salt, either KI (2.0 mg, 0.012 mmol) or LiBr (1.03 mg, 0.012 mmol).

The NMR data for the following compounds has been deposited in the Natural Products Magnetic Resonance Database (NP-MRD; www.np-mrd.org) and can be found at NP0351618 (Tartrolon E Lithium Salt) and NP0351620 (Tartrolon E Potassium Salt). Crystallography data has been deposited in the Cambridge Crystallographic Data Centre under deposition number 2473545 for the structure of the tartrolon E-Na+ complex.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jnatprod.5c01073.

• Additional experimental data from fermentation experiments, 1D and 2D NMR spectra, and crystallographic data­(PDF)

This work was funded by the National Institutes of Health Awards NIH R01AI162943 and NIH R01AI127724. NMR results included in this report were recorded at the David M. Grant Center, a University of Utah Core Facility, and at the University of Utah Health Sciences Center NMR Core Facility. Funds for construction of the Grant Center and the helium recovery system were obtained from the University of Utah and the National Institutes of Health awards 1C06RR017539-01A1 and 3R01GM063540-17W1, respectively. Grant Center NMR instruments were purchased with support of the University of Utah and the National Institutes of Health award 1S10OD25241-01. Crystallography reported in this publication was supported by the Office of the Director, National Institutes of Health under Award S10OD030326. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare the following competing financial interest(s): E.W.S. is a co-inventor on a patent applying tartrolons to parasitic infections.