Solid-State Nuclear Magnetic Resonance Analysis Reveals a Possible

Dec 28, 2016 - In this work, to examine the Ca2+ binding site of PRM-A, we performed a solid-state nuclear magnetic resonance experiment using 111Cd2+...
32 downloads 10 Views 610KB Size
Download PDF
Article pubs.acs.org/biochemistry

Solid-State Nuclear Magnetic Resonance Analysis Reveals a Possible Calcium Binding Site of Pradimicin A Takashi Doi,† Yu Nakagawa,‡ and K. Takegoshi*,† †

Department of Chemistry, Graduate School of Science, Kyoto University, Kitashirakawa Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan ‡ Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan S Supporting Information *

ABSTRACT: Pradimicin A (PRM-A) is a unique natural product that recognizes D-mannopyranoside (Man) in the presence of Ca2+ ion. Although the Man binding geometry of PRM-A is largely understood, the molecular basis of Man recognition has yet to be established because of the lack of information regarding Ca2+ binding geometry. In this work, to examine the Ca2+ binding site of PRM-A, we performed a solidstate nuclear magnetic resonance experiment using 111Cd2+ as a surrogate probe for Ca2+. Evaluation of 13C−111Cd distances in the [PRM-A/111Cd2+] complexes by rotational-echo double resonance (REDOR) and 111Cd frequency selective REDOR (FSR) revealed that PRM-A binds 111Cd2+ at the anthraquinone moiety, which contradicts the previous hypothesis of the alanine moiety being the Ca2+ and Cd2+ binding sites of PRM-A. The distances between Cd2+ and the carbon atoms at the binding site of PRM-A were found to be 3.5 ± 0.2 Å. Importantly, Man binding was shown not to alter the distances, indicating that [PRM-A/Ca2+] and [PRM-A/Ca2+/Man] complexes have similar Ca2+ binding geometries. This study provides an important clue to understanding the molecular basis of Man recognition of PRM-A.

P

PRM-A recognizes Man in aqueous solutions should contribute to the design of the low-molecular weight CBMs. However, structural determination of the complex of PRM-A with Ca2+ and Man has not been fully achieved because of the difficulties described below. The complexes of PRM-A with Ca2+ and Man have three forms in equilibrium. A pair of PRM-A and a Ca2+ ion initially forms a binary [PRM-A2/Ca2+] complex, which binds two more Man molecules and makes the ternary [PRM-A2/Ca2+/Man2] complex. It changes into the ternary [PRM-A2/Ca2+/Man4] complex with two additional Man molecules.9 Because these three forms of complexes precipitate to make complicated aggregates, application of a conventional structural determination tool such as X-ray diffraction is difficult. Recently, we showed that intermolecular 13C−13C distances between PRM-A and methyl α-D-mannopyranoside (Man-OMe) in the 13Cenriched ternary [PRM-A2/Ca2+/Man-OMe2] complex can be obtained by using solid-state NMR techniques.10−12 These 13 C−13C distance constraints will give an overall structure of the ternary PRM-A complex; however, the Ca2+ binding site has not yet been specified. Thus far, the alanine moiety of PRM-A is considered as a putative Ca2+ binding site because its esterification abolishes the

radimicin A [PRM-A (Figure 1)] is a natural lowmolecular weight nonpeptidic antibiotic that recognizes

Figure 1. Representative structure of pradimicin A (PRM-A). Gray circles indicate 13C-enriched positions of [13C12]PRM-A. A black circle indicates a 13C-enriched position of [18-13C]PRM-A. The number of a 13 C-enriched carbon is also shown.

(Man) specifically in the presence of Ca2+ ion in aqueous solutions.1−3 This unique Man recognition ability allows PRM-A to exhibit antifungal and anti-HIV properties.4,5 PRM-A also has attracted considerable interest in the context of artificial small-size carbohydrate-binding molecules (CBM) used to mimic lectin.6−8 Disclosing how D-mannopyranoside

© 2016 American Chemical Society

Received: December 25, 2016 Published: December 28, 2016 468

DOI: 10.1021/acs.biochem.6b01300 Biochemistry 2017, 56, 468−472

Article

Biochemistry Ca2+ binding ability of PRM-A.13 However, some experimental results contradict the assumption of the alanine moiety coordinating with Ca2+. As an example, Nishio and coworkers14 reported that the amide derivatives of PRM-A also showed significant antifungal activities in the presence of Ca2+ ion and the alanine moiety could be modified while their antifungal activities were maintained. In this work, we determine the Ca2+ binding site of PRM-A by using solidstate NMR. Solid-state NMR is widely utilized to obtain structural insights into metal-coordinated biomolecules and metalloproteins. For example, Pourpoint and co-workers15 have recently reported RFDR experiments for 13C−51V distance measurement in the oxovanadium(V) complex. In principle, distances between the carbons of PRM-A and the Ca2+ ion and thereby the position of the Ca2+ ion in the complex can be determined by 13C−43Ca double-resonance solid-state NMR. For example, it was shown that 13C−43Ca distances in the calcium benzoate trihydrate were determined by using 13 C−43Ca transfer of population double resonance (TRAPDOR).16 In practice, however, such an experiment is difficult because of the low natural abundance (0.15%) and the small gyromagnetic ratio (γCa/γH = 0.067) of 43Ca. In addition, 43Ca NMR becomes much difficult if the quadrupolar interaction is strong. In this work, we use Cd2+ as a surrogate probe for Ca2+ and apply 13C−111Cd double-resonance solid-state NMR to measure C−Cd distances in the PRM-A complexes with Cd2+. There have been a number of 113Cd NMR works that have examined the Ca2+ binding sites in metalloproteins by using 113Cd2+ as a surrogate probe for Ca2+.17,18 Considering that the difference in Larmor frequencies between 13C and Cd is larger for 111Cd, we chose 111Cd in this work. The 13C−111Cd distance can be measured simply by using rotational-echo double resonance (REDOR)19 because the spin quantum number of 111Cd is 1/2. High-power decoupling on 1H spins is also desirable. To perform the 1H−13C−111Cd experiment, we modified a Chemagnetics commercial T3 triply tuned NMR probe by adding a special frequency trap to isolate 13C and 111 Cd channels. As for the sample, we prepared the complexes of 13C-enriched PRM-A with 111Cd2+ and 25 equiv of ManOMe to PRM-A, for which we showed in the previous Cd NMR study that the sample is a mixture of the binary and ternary complexes. In this work, we determine the Ca (111Cd) binding site in both the binary and the ternary complexes by applying 111Cd frequency selective REDOR (FSR)20 to the two 111 Cd signals of the binary [PRM-A2/111Cd2+] complex and the ternary [PRM-A2/111Cd2+/Man-OMe2] complex separately. From the 13C−111Cd internuclear distances determined by REDOR and FSR, we conclude that the Cd binding site is not the alanine moiety but the anthraquinone moiety of PRM-A, and the 13C−111Cd distances for the carbons at the binding site are 3.5 ± 0.2 Å. The 13C−111Cd distance measurement demonstrated in this work by using Cd2+ as a surrogate for a metal ion should be generally applicable for structural determination of a metal-binding biomolecule, which is prone to aggregate to hamper the X-ray crystallographic or solution NMR analysis.

the intramolecular 13C−13C dipolar interactions in the REDOR experiments, the 13C enrichment ratio in [13C12]PRM-A was set to ∼20 atom %, which was confirmed by solution 1H NMR.11 In the complex of [18-13C]PRM-A, C18, the carboxyl carbon of the alanine moiety is ∼70 atom % enriched with 13C. The Cd2+ ion in the complex is fully enriched with 111Cd; with these enrichment patterns, we were able to observe 13C−111Cd distances by using the FSR experiment with 13C being the observed one. Detailed preparation procedures of the complexes of PRM-A are described in the Supporting Information. Solid-State NMR Experiments. All solid-state NMR experiments were performed with a 9.4 T magnet (JASTEC) using an OPENCORE spectrometer21 and a 1H−13C−111Cd triply tuned probe (Chemagnetics) with a home-built trap plugin. A 3.2 mm rotor was used with the magic angle spinning (MAS) frequency being controlled with a home-built spinning frequency controller within ±5 Hz. Detailed experimental parameters of solid-state NMR are described in the Supporting Information.

■

RESULTS AND DISCUSSION For 13C−111Cd distance measurements, we prepared solid powder PRM-A samples with 111Cd2+ and 25 equiv of ManOMe to PRM-A. This led a mixture of the binary and ternary complexes. The ratio of the ternary complex to the binary one was evaluated from the two 111Cd signals in the 111Cd onedimensional (1D) Hahn-echo spectrum (Figure 2). The

Figure 2. 111Cd 1D NMR spectrum of the complex of [13C12]PRM-A with 111Cd2+. The sharper signal (δ −135) is of the ternary complex, and the broader signal (δ −50) is of the binary complex. The ratio of the ternary complex to the binary one obtained by a line shape fitting is ∼2:1 (fitted curves not shown).

spectrum is similar to the previously reported one for the PRM-A complex prepared with 113Cd2+ and 25 equiv of ManOMe to PRM-A.10 We assigned the sharper signal at δ −135 to the ternary complex and the broader one around δ −50 to the binary complex. From the ratio of the area of the δ −135 signal to that of the δ −50 signal, the ratio of the ternary complex to the binary one was estimated to be ∼2:1. First, we examined the conventional nonselective REDOR experiment. Figure 3 shows the 13C REDOR reference [S0 (Figure 3a)] and REDOR difference [ΔS = S0 − SR (Figure 3b)] spectra of the complex of [13C12]PRM-A at a dipolar dephasing time of 9.76 ms; the signal intensities of C13 and C14 significantly decrease. The number of a carbon is shown in Figure 1. Note that the 13C signals for the binary and ternary complexes are not resolved. Therefore, the observed ΔS for a given carbon is a sum of the two 13C signals in the two complexes and may be written as

■

MATERIALS AND METHODS Preparation of 13C-Enriched PRM-As. 13C-enriched PRM-As were prepared as reported previously.11 To ignore 469

DOI: 10.1021/acs.biochem.6b01300 Biochemistry 2017, 56, 468−472

Article

Biochemistry

we would like to measure distances among one particular 111Cd spin and several 13C spins, we replaced a Gaussian π pulse for 13 C in the original FSR with a hard π pulse, while a Gaussian π pulse is applied to 111Cd spins (Figure S3b). Note that the possible effects of the 13C−13C dipolar interactions on the ΔS/ S0 curve are negligible as the 13C enrichment ratio in [13C12]PRM-A is only ∼20 atom %. Furthermore, all rotorsynchronized π pulses were applied to 111Cd spins to avoid the recoupling of 13C spins.22,23 Upon application of the frequency selective π pulse on resonance to, say, the δ −50 111Cd signal, the resulting FSR difference signal for a given carbon may be expressed by setting a1 ∼ 1.0 and a2 ∼ 0.0 in eq 1:

Figure 3. (a) 13C 1D nonselective REDOR reference (S0) spectrum and (b) difference (ΔS) spectrum of the complex of [13C12]PRM-A with 111Cd2+ at a dephasing time of 9.76 ms. Peak assignments are shown in Figure S2.

ΔS = fa1ΔS1 + (1 − f )a 2ΔS2

ΔS obs = 0.33ΔS1

Similarly, by selective inversion of the δ −135 signal, we have

ΔS obs = 0.67ΔS2

(1)

Figure 5 shows the observed dipolar dephasing of the C13/C14 signals with the two selective-inversion conditions. For

where f is the ratio of the binary complex, ai is the inversion ratio of the 111Cd signals by the π pulse, and ΔSi is the REDOR difference signal intensity (i = 1 and 2 denote the binary complex and the ternary complex, respectively). In the sample presented here, the ratio f is estimated to be ∼0.33 from the 111 Cd signal intensities. We assume both of the inversion ratios ai to be 1.0; i.e., a perfect inversion is achieved. The dephasing time dependence of ΔS/S0 values is plotted in Figure 4. At the longer dephasing time, the ΔS/S0 values of

Figure 4. Dephasing time dependence of the ΔS/S0 values obtained from the nonselective REDOR experiment. Error bars were calculated like those given in Figure S4. Solid lines are guides for the eye.

Figure 5. Dephasing time dependence of (a) the C13 ΔS/S0 values and (b) the C14 ΔS/S0 values. The ΔS/S0 values obtained from the nonselective REDOR experiment (circles), those obtained from the FSR experiment with the Gaussian π pulse applied to the δ −135 signal of 111Cd (squares), and those obtained from the FSR experiment for the δ −50 signal of 111Cd (triangles) are plotted. Error bars were calculated like those shown in Figure S4. The black solid lines in both plots are the calculated curves with a 13C−111Cd internuclear distance of 3.5 Å. The dotted lines are the calculated curves for 3.3 Å and the broken lines for 3.7 Å.

C13 and C14 increase significantly despite those of other carbons of PRM-A, such as C2 and C3-Me, remaining rather constant. These data unambiguously show that the Cd2+ ion is close to C13/C14 of PRM-A. It is also notable that the ΔS/S0 of C7a increases moderately. This increase is consistent because C7a is a next neighbor carbon of C13/C14. It should be noted further that the dephasing curve of C18, which was obtained from another REDOR experiment using the complex of [18-13C]PRM-A with 111Cd2+, indicates a long 13C−111Cd distance. These results clearly indicate that the Cd binding site of PRM-A is the anthraquinone moiety including C13/C14 and not the alanine moiety including C18. Having shown qualitatively that the alanine moiety is not the Cd binding site by conventional nonselective REDOR, we determined the distances between C13/C14 and 111Cd quantitatively. This was done by applying 111Cd selective FSR to the complex of [13C12]PRM-A with 111Cd2+. In the original FSR experiment, a pair of Gaussian pulses is applied simultaneously to both I spins and S spins to select one I−S spin pair in a multiply I/S isotopically labeled sample. Because

comparison, the results for the nonselective experiments (Figure 4) are also plotted. Both of the ΔS/S0 curves for the ternary complex with its 111Cd signal at δ −135 show clear REDOR dipolar dephasing, which are consistent with the curves calculated for a 13C−111Cd distance of 3.5 Å. We thus concluded that the C13−Cd and C14−Cd distances are 3.5 ± 0.2 Å in the ternary complex. Because of the small ratio ( f ∼ 0.33) of the binary complex, the amplitudes of the ΔS/S0 curves are reduced. In our observation, the sum of ΔS/S0 values of FSR experiments is slightly smaller than the ΔS/S0 in the REDOR experiment (Figure S8). We consider it is due to an imperfect inversion of the selective Gaussian π pulses; it is 470

DOI: 10.1021/acs.biochem.6b01300 Biochemistry 2017, 56, 468−472

Article

Biochemistry difficult to invert the considerably broad 111Cd signals perfectly by the Gaussian π pulse with the limited pulse duration time. The imperfect inversion by the Gaussian π pulse (ai < 1.0) may also lead the damping of the amplitude and induce slightly small signal intensities relative to conventional REDOR. However, even with these reservations, the C13−Cd and C14−Cd distances in the binary complex are similar to those in the ternary complex. With the information about the Cd binding site being at C13/C14, we reexamined the 13C spectra of the ternary complex of [13C12]PRM-A with 111Cd2+ and that with Ca2+ to find an ∼1 ppm shift of the C13/C14 signals (Figure S2 and Table S1). These shifts can be ascribed to the change in the coordinated metal ion and are supporting evidence of C13/C14 being the Ca binding site. In this work, we showed that the Cd2+ binding site of PRM-A is not the alanine moiety but the anthraquinone moiety of PRM-A. Even in the presence of Cd2+, the unique lectin-like property of PRM-A to recognize Man-OMe is preserved; we consider that the Ca binding site is also at the C13/C14 site in the anthraquinone moiety, the same as the Cd binding site in the complex of PRM-A with Cd2+. Also, the Ca2+ ion has been considered to act as a bridge to bind PRM-A to Man-OMe. The Ca2+ coordination of Man-OMe is consistent with the proximity of Man-OMe to PRM-A previously reported by using 13C−13C distance geometry.10−12

Funding

This research was partly supported by a SUNBOR GRANT from the Suntory Foundation of Life Sciences and a MEXT Grant-in Aid for Scientific Research (B) (Grant 15H04496). Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS We thank Prof. Yukishige Ito and Prof. Yasuhiro Igarashi for their generous support. ABBREVIATIONS CBM, carbohydrate-binding molecule; FSR, frequency selective REDOR; HIV, human immunodeficiency virus; Man, Dmannopyranoside; Man-OMe, methyl α-D-mannopyranoside; MAS, magic angle spinning; NMR, nuclear magnetic resonance; PRM-A, pradimicin A; REDOR, rotational-echo double resonance; TRAPDOR, transfer of population double resonance.

■

(1) Oki, T., Konishi, M., Tomatsu, K., Tomita, K., Saitoh, K., Tsunakawa, M., Nishio, M., Miyaki, T., and Kawaguchi, H. (1988) Pradimicin, a novel class of potent antifungal antibiotics. J. Antibiot. 41, 1701−1704. (2) Takeuchi, T., Hara, T., Naganawa, H., Okada, M., Hamada, M., Umezawa, H., Gomi, S., Sezaki, M., and Kondo, S. (1988) New antifungal antibiotics, benanomicins A and B from an actinomycete. J. Antibiot. 41, 807−811. (3) Sawada, Y., Numata, K., Murakami, T., Tanimichi, H., Yamamoto, S., and Oki, T. (1990) Calcium-dependent anticandidal action of pradimicin A. J. Antibiot. 43, 715−721. (4) Balzarini, J. (2007) Targeting the glycans of glycoproteins: a novel paradigm for antiviral therapy. Nat. Rev. Microbiol. 5, 583−597. (5) Balzarini, J., Van Laethem, K., Daelemans, D., Hatse, S., Bugatti, A., Rusnati, M., Igarashi, Y., Oki, T., and Schols, D. (2007) Pradimicin A, a carbohydrate-binding nonpeptidic lead compound for treatment of infections with viruses with highly glycosylated envelopes, such as human immunodeficiency virus. J. Virol. 81, 362−373. (6) Walker, D. B., Joshi, G., and Davis, A. P. (2009) Progress in biomimetic carbohydrate recognition. Cell. Mol. Life Sci. 66, 3177− 3191. (7) Mazik, M. (2009) Molecular recognition of carbohydrates by acyclic receptors employing noncovalent interactions. Chem. Soc. Rev. 38, 935−956. (8) Jin, S., Cheng, Y., Reid, S., Li, M., and Wang, B. (2010) Carbohydrate recognition by boronolectins, small molecules, and lectins. Med. Res. Rev. 30, 171−257. (9) Fujikawa, K., Tsukamoto, Y., Oki, T., and Chuan Lee, Y. (1998) Spectroscopic studies on the interaction of pradimicin BMY-28864 with mannose derivatives. Glycobiology 8, 407−414. (10) Nakagawa, Y., Masuda, Y., Yamada, K., Doi, T., Takegoshi, K., Igarashi, Y., and Ito, Y. (2011) Solid-state NMR spectroscopic analysis of the Ca2+-dependent mannose binding of pradimicin A. Angew. Chem., Int. Ed. 50, 6084−6088. (11) Nakagawa, Y., Doi, T., Masuda, Y., Takegoshi, K., Igarashi, Y., and Ito, Y. (2011) Mapping of the primary mannose binding site of pradimicin A. J. Am. Chem. Soc. 133, 17485−17493. (12) Nakagawa, Y., Doi, T., Taketani, T., Takegoshi, K., Igarashi, Y., and Ito, Y. (2013) Mannose-binding geometry of pradimicin A. Chem. - Eur. J. 19, 10516−10525. (13) Ueki, T., Numata, K., Sawada, Y., Nishio, M., Ohkuma, H., Toda, S., Kamachi, H., Fukagawa, Y., and Oki, T. (1993) Studies on the mode of antifungal action of pradimicin antibiotics II. Dmannopyranoside-binding site and calcium-binding site. J. Antibiot. 46, 455−464.

■

CONCLUSION We successfully showed in this work that the 111Cd binding site of PRM-A is not the carboxyl group of the alanine moiety, which has been considered as a binding site, but the C13/C14 site in the anthraquinone moiety by 13C−111Cd distance measurements using solid-state NMR. The C−Cd distances between C13/C14 of PRM-A and the Cd2+ ion are found to be 3.5 ± 0.2 Å in the binary and ternary complexes. Because PRMA maintains the ability to recognize Man selectively in the presence of Cd2+, it is natural to consider that the Ca binding site is at the C13/C14 site in the anthraquinone moiety of PRM-A. This consideration is also confirmed by the metalinduced shifts of the C13/C14 signals in the complexes of [13C12]PRM-A with 111Cd2+ or Ca2+. At present, examination of the Man-OMe binding geometry with the Cd2+ ion using the REDOR approach is being undertaken with a hope of disclosing the overall structure of the ternary complex, which will give us perspective on exploiting novel anti-HIV drugs and designing conceptually new CBMs working in aqueous solutions.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.6b01300. Detailed information about sample preparation, detailed experimental parameters of solid-state NMR, additional NMR spectra, diagrams of the pulse sequences, and additional REDOR and FSR plots (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 471

DOI: 10.1021/acs.biochem.6b01300 Biochemistry 2017, 56, 468−472

Article

Biochemistry (14) Nishio, M., Ohkuma, H., Kakushima, M., Ohta, S., Iimura, S., Hirano, M., Konishi, M., and Oki, T. (1993) Synthesis and antifungal activities of pradimicin A derivatives modification of the alanine moiety. J. Antibiot. 46, 494−499. (15) Pourpoint, F., Yehl, J., Li, M., Gupta, R., Trébosc, J., Lafon, O., Amoureux, J. P., and Polenova, T. (2015) NMR crystallography of an oxovanadium(V) complex by an approach combining multinuclear magic angle spinning NMR, DFT, and spin dynamics simulations. ChemPhysChem 16, 1619−1626. (16) Laurencin, D., Gervais, C., Wong, A., Coelho, C., Mauri, F., Massiot, D., Smith, M. E., and Bonhomme, C. (2009) Implementation of high resolution 43Ca solid state NMR spectroscopy: toward the elucidation of calcium sites in biological materials. J. Am. Chem. Soc. 131, 13430−13440. (17) Coleman, J. E. (1993) Cadmium-113 nuclear magnetic resonance applied to metalloproteins. Methods Enzymol. 227, 16−43. (18) Clarke, T. E., and Vogel, H. J. (1992) Cadmium-113 and Lead207 NMR spectroscopic studies of calcium-binding proteins. Methods Mol. Biol. 173, 205−215. (19) Gullion, T., and Schaefer, J. (1989) Rotational-echo doubleresonance NMR. J. Magn. Reson. 81, 196−200. (20) Jaroniec, C. P., Tounge, B. A., Herzfeld, J., and Griffin, R. G. (2001) Frequency selective heteronuclear dipolar recoupling in rotating solids: accurate 13C-15N distance measurements in uniformly 13 C, 15N-labeled peptides. J. Am. Chem. Soc. 123, 3507−3519. (21) Takeda, K. (2008) OPENCORE NMR: Open-source core modules for implementing an integrated FPGA-based NMR spectrometer. J. Magn. Reson. 192, 218−229. (22) Gullion, T., and Vega, S. (1992) A simple magic angle spinning NMR experiment for the dephasing of rotational echoes of dipolar coupled homonuclear spin pairs. Chem. Phys. Lett. 194, 423−428. (23) Bennett, A. E., Ok, J. H., Griffin, R. G., and Vega, S. (1992) Chemical shift correlation spectroscopy in rotating solids: radio frequency−driven dipolar recoupling and longitudinal exchange. J. Chem. Phys. 96, 8624−8627.

472

DOI: 10.1021/acs.biochem.6b01300 Biochemistry 2017, 56, 468−472