Binding of a Fatty Acid-Functionalized Anderson-Type Polyoxometalate to Human Serum Albumin

The Anderson-type hexamolybdoaluminate functionalized with lauric acid (LA), (TBA)3[Al(OH)3Mo6O18{(OCH2)3CNHCOC11H23}]·9H2O (TBA-AlMo6-LA, where TBA = tetrabutylammonium), was prepared via two synthetic routes and characterized by thermogravimetric and elemental analyses, mass spectrometry, IR and 1H NMR spectroscopy, and powder and single-crystal X-ray diffraction. The interaction of TBA-AlMo6-LA with human serum albumin (HSA) was investigated via fluorescence and circular dichroism spectroscopy. The results revealed that TBA-AlMo6-LA binds strongly to HSA (63% quenching at an HSA/TBA-AlMo6-LA ratio of 1:1), exhibiting static quenching. In contrast to TBA-AlMo6-LA, the nonfunctionalized polyoxometalate, Na3(H2O)6[Al(OH)6Mo6O18]·2H2O (AlMo6), showed weak binding toward HSA (22% quenching at a HSA/AlMo6 ratio of 1:25). HSA binding was confirmed by X-ray structure analysis of the HSA-Myr-AlMo6-LA complex (Myr = myristate). These results provide a promising lead for the design of novel polyoxometalate-based hybrids that are able to exploit HSA as a delivery vehicle to improve their pharmacokinetics and bioactivity.


Experimental Section
All starting materials were purchased from Sigma-Aldrich, TCI, or Acros and used as received unless otherwise specified. Fatty acid and globulin free HSA from Sigma-Aldrich (A3782) was used for the spectroscopic experiments (without further purification). The educts Na 3 (H 2 O) 6 [Al(OH) 6 Mo 6 O 18 ]·2H 2 O (AlMo 6 ) and (TBA) 3 [Al(OH) 3 Mo 6 O 18 {(OCH 2 ) 3 CNH 2 }]·7H 2 O (TBA-AlMo 6 -Tris, TBA = tetrabutylammonium) were prepared according to reported procedures. 1,2 Powder X-ray diffraction was performed on an EMPYREAN diffractometer system using Cu Kα radiation (λ= 1.540598), a PIXcel3D-Medipix3 1x1 detector (used as a scanning line detector) and a divergence slit fixed at 0.1 mm. The scan range was from 5° to 50°(2θ). 1 H-NMR spectra were recorded on Bruker Avance III spectrometers at 500.32 and 500.10 MHz, respectively. 1 H shifts are quoted relative to the solvent residual signals. Electrospray ionization (ESI) mass spectra were measured on a Bruker Esquire 3000 mass spectrometer. The measurement was carried out in methanol, whereby the data were collected in positive mode within the region of m/z 100-1200. The m/z values are quoted for the most abundant isotope. Infrared (IR) spectra were obtained from a Bruker Vertex 70 FT-IR spectrometer by means of the attenuated-total-reflection (ATR) technique. All elemental analyses were carried out using an `EA 1108 CHNS-O' elemental analyzer (Carlo Erba Instruments) at the Microanalytical Laboratory of the University of Vienna. Thermogravimetric analysis (TGA) was performed on a Mettler SDTA851e Thermogravimetric Analyzer under nitrogen flow, with a heating rate of 5 K/min in the region of 298-473 K. We performed a literature search using the SCOPUS (https://www.scopus.com/) and CCDC (https://www.ccdc.cam.ac.uk/structures/) databases to identify POMs functionalized by fatty acids, or more precisely, hybrid POMs containing aliphatic chains (≥ C3 atoms). The results of the literature search are summarized in Table S1.

Route 2:
0.26 ml of triethylamine (1.7 mmol) was added to a solution of (TBA) 3 [Al(OH) 3 Mo 6 O 18 (OCH 2 ) 3 CNH 2 ]·7H 2 O (2.7 g, 1.5 mmol) in 20 ml of anhydrous acetonitrile. Afterward, 1.7 mmol lauroyl chloride (0.37 g) was added dropwise to that solution. The mixture was stirred at room temperature for 48 hours, filtered and evaporated under vacuum. The residue was washed with THF and water and recrystallized from a methanol-water mixture (1:1). Yield: 2.5 g, 77 %. TGA (see Figure S1) revealed that 7.71 % of the weight is lost when the compound is heated from 25 -200 °C, which corresponds to nine water molecules.
IR and 1 H-NMR spectra are identical to those of the product obtained by route 1. Na-AlMo 6 -LA was obtained from TBA-AlMo 6 -LA via ion-exchange using a column with Amberlite IR-120 in Na + -form and 50 % methanol as solvent.
Anal. Calcd for Na 3

Crystallization of HSA-Myr
Crystallization of HSA was only achieved in the presence of myristate (Myr) because the fatty acid-induced conformational changes facilitate the crystallization process of the protein. For this reason, HSA was dissolved in 50 mM potassium phosphate, pH 7.4, and 150 mM KCl, and incubated with a 10-fold excess of sodium myristate for 3 hours at 37 °C. Unbound myristate was removed by centrifugation at 4000 rpm for 15 min (at 4 °C). HSA-Myr was concentrated to 150 mg/ml in the above-mentioned buffer and subjected to crystallization. Initial crystallization trials were performed by the hanging-drop vapor-diffusion technique using a 15-well Easy-Xtal plate (Qiagen). Several microliters of the protein solution (100 -200 mg/ml HSA-Myr) were mixed with 2 µl of reservoir solution (50 mM potassium phosphate, pH 7.5 -8.0, 23-30 % PEG3350, 150 mM KCl) and incubated at 293 K. Initial crystals grew, however, in clusters and were highly twinned. Therefore, streak seeding was performed in drops that had been allowed to equilibrate for three days. Crystals continued to grow in clusters (even after three rounds of streak seeding) but were clearly less twinned. Thus, it was possible to detach single crystals for further crystallization experiments.

Crystallization of TBA-AlMo 6 -LA
Initial crystals of TBA-AlMo 6 -LA were obtained by redissolving the solid (solubility ~2 mM) in water and crystallization on air. However, the crystals were of relatively low quality, leading to incomplete structures during the subsequent structure solving process. High-quality crystals were obtained during the attempt to co-crystallize TBA-AlMo 6 -LA with HSA-Myr. To crystallize the complex, the hanging-drop vapor-diffusion technique was applied using a 15-well Easy-Xtal plate (Qiagen). Single crystals of TBA-AlMo 6 -LA grew at 293 K by mixing several microliters of protein solution (100 -200 mg/ml HSA-Myr) with 2 µl of reservoir solution (50 mM potassium phosphate, pH 7.5 -8.0, 23-30 % PEG3350, 150 mM KCl) and 2 µL of TBA-AlMo 6 -LA (~2 mM). Crystals of TBA-AlMo 6 -LA appeared within 24 hours. Interestingly, the presence of HSA was essential for the formation of high-quality crystals of AlMo 6 -LA because in the absence of the protein only small needles were obtained that were not suitable for X-ray diffraction.

X-ray diffraction experiment and structure elucidation of TBA-AlMo 6 -LA
Crystals of TBA-AlMo 6 -LA were mounted in nylon loops and flash-cooled in liquid nitrogen after a quick soak in a cryoprotectant solution (50 mM potassium phosphate, pH 7.5 -8.0, 32 % PEG3350 and 150 mM KCl). X-ray data of TBA-AlMo 6 -LA were measured on a Bruker D8 Venture diffractometer equipped with a multilayer monochromator, Cu K/α INCOATEC microfocus sealed tube, and an Oxford cooling system. The structure was solved by direct methods and refined by full-matrix least-squares techniques. Non-hydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were inserted at calculated positions and refined with the riding model; only the three hydrogen atoms at the POM core were refined free with the use of restraints. The following programs were used: Bruker SAINT software package 3 using a narrow-frame algorithm for frame integration; SADABS 4 for absorption correction; OLEX2 5 for structure solution, refinement, and molecular diagrams; graphical user-interface Shelxle 6 for refinement; graphical user-interface SHELXS-2015 7 for structure solution; SHELXL-2015 8 for refinement; Platon 9 for symmetry check. Experimental data with the CCDC-Code (1944363) are available online: http://www.ccdc.cam.ac.uk/conts/retrieving.html. Crystal data, data collection parameters, and structure refinement details are given in Tables S2. Selected bond lengths are summarized in Table S3.

Fluorescence spectroscopy
Fluorescence emission spectra of HSA (5 µM) in absence and presence of various concentrations of TBA-AlMo 6 -LA, AlMo 6 , lauric acid, TBA, and Na-AlMo 6 -LA (1.25, 2.5, 5, 10, 25, 50 and 125 µM) were measured on an Infinite 200 microplate reader (Tecan) using Microfluor 1Black flat-bottom microtiter plates (Thermo Scientific). Fluorescence measurements were performed at least in triplicates and at three different temperatures: 303, 308, and 310.5 K. The excitation and emission wavelengths were set at 270 nm and 320-450 nm, respectively. All samples were dissolved in 20 mM potassium phosphate (pH 7.4) and 150 mM KCl. Because TBA/Na-AlMo 6 -LA and AlMo 6 exhibit substantial absorption at the excitation wavelength (and partially at the emission wavelengths) at high concentrations, the obtained fluorescence intensities had to be corrected for the inner filter effect. To account for the inner filter effect, the following formula was used: 10 where F corr and F obs are the fluorescence intensities after and before correcting the inner filter effect, respectively, OD ex and OD em are the optical density at the excitation and emission wavelengths, respectively, and 0.605 is a factor considering the geometry of the plate wells.
The quenching effect of all quenchers (AlMo 6 -LA, AlMo 6 , and lauric acid) was evaluated by both the common and modified Stern-Volmer equations (see main text). The fluorescence spectra of HSA with AlMo 6 -LA, AlMo 6, and lauric acid, respectively, are shown in the Figures S5-S8 and Figure 2 (main text).

CD spectroscopy
CD spectra were collected using a Chirascan Plus spectropolarimeter (Applied Photophysics). The apparatus was sufficiently purged with nitrogen gas before starting the experiment. All CD measurements were carried out at 303 K using a CS/PCS single cell controller. Each experiment was performed using a precision quartz cuvette of 0.1 cm pathlength, and data were collected at wavelengths between 200 and 260 nm. Every spectrum shown here is the average of five successive scans, whereby the data were baseline subtracted for buffer. All samples were prepared in 20 mM potassium phosphate buffer at pH 7.5 with 150 mM KCl and a constant HSA concentration of 0.25 mg/ml (3.76 µM). The concentration of TBA-AlMo 6 -LA, AlMo 6 , and lauric acid, respectively, varied from 3.76 to 14 µM. The α-helicity of free HSA and the HSA-AlMo 6 -LA-, HSA-AlMo 6 -and HSA-LA-complexes was calculated using the following equation 11 : where Θ 222 is the observed mean residue ellipticity at 222 nm (in deg cm 2 /dmol) and 36000 and 3000 are the Θ values of a pure α-helix, β-sheet and random coil conformation at 222 nm, respectively. Θ 222 was calculated from each spectrum using the following equation 12 : where θ obs is the measured ellipticity in millidegrees (mdeg), n is the number of amino acids in HSA, C p is the molar concentration of HSA and l is the pathlength of light in cm of the cuvette. The α-helicity of each sample is summarized in Table S5.

Crystallization of HSA-Myr and subsequent soaking with TBA-AlMo 6 -LA
HSA-Myr crystals were obtained as described in section 1.4. and transferred into stabilizing drops containing 50 mM potassium phosphate, pH 7.5 -8.0, 35 % PEG3350, 150 mM KCl, and 2 mM Myr. Soaking of HSA-Myr crystals in TBA-AlMo 6 -LA-containing drops did not yield crystals of the complex because of the low solubility of the hybrid POM (~ 2 mM). However, AlMo 6 -LA was successfully introduced into the HSA structure by directly adding TBA-AlMo 6 -LA powder into the stabilizing drop. After soaking overnight (longer soaking times had negative effects on the crystals), crystals were harvested in nylon loops, quickly wiped through a freshly prepared stabilizing drop to remove POM powder, and subsequently flash-frozen in liquid nitrogen. The complex was successfully formed in four of five trials; however, crystal soaking with unmodified AlMo 6 (using the same procedure) did not yield the complex structure (in four trials).
1.10. X-ray diffraction experiment and structure elucidation of HSA-Myr-AlMo 6 -LA X-ray data of the HSA-Myr-AlMo 6 -LA complex were collected at 100 K using a BRUKER D8 VENTURE X-ray diffractometer equipped with a multilayer monochromator, a PHOTON II charge-integrating pixel array detector, and a Cu-Kα Incoatec Microfocus (sealed tube). Diffraction data were processed with XDS 13 , whereby two datasets of the same crystal were merged with XSCALE 14 to obtain one dataset with reasonable completeness. Initial phases were obtained by the molecular replacement method applying PHASER 15 ; PDB entry 1N5U 16 was used as a search model. The obtained structure was refined with phenix.refine 17 and occasionally manually edited with Coot 18 . After the refinement has reached convergence, two molecules of AlMo 6 -LA (without TBA cations) were introduced into the structure with COOT. Coordinates and restraints for the POM were generated with phenix.elbow 19 using the coordination file of TBA-AlMo 6 -LA (CCDC entry 1944363) as input. The HSA-Myr-AlMo 6 -LA structure was further refined until convergence, whereby the occupancies and the B-factors (using anisotropic ADPs for the metals) were separately refined for the POMs. Data collection and refinement statistics are summarized in Table S6. Figure S1. TGA curve of TBA-AlMo 6 -LA. The TGA curve shows a two-step weight loss process (difference plot is shown in red). The weight loss during the first step was due to dehydration (25-200 °C), where 7.71 % of the weight was lost, corresponding to nine lattice water molecules. During the second step, 1.03 % of the weight was lost, which might be due to AlMo 6 -LA decomposition. Please note that the maximum temperature of 200 °C was reached after 45 minutes and was then kept constant until the end of the experiment. 8 Figure S1. TGA curve of TBA-AlMo 6 -LA. The TGA curve shows a two-step weight loss process (difference plot is shown in red). The weight loss during the first step was due to dehydration (25-200 °C), where 7.71 % of the weight was lost, corresponding to nine lattice water molecules. During the second step, 1.03 % of the weight was lost, which might be due to AlMo 6 -LA decomposition. Please note that the maximum temperature of 200 °C was reached after 45 minutes and was then kept constant until the end of the experiment. 8 Figure S1. TGA curve of TBA-AlMo 6 -LA. The TGA curve shows a two-step weight loss process (difference plot is shown in red). The weight loss during the first step was due to dehydration (25-200 °C), where 7.71 % of the weight was lost, corresponding to nine lattice water molecules. During the second step, 1.03 % of the weight was lost, which might be due to AlMo 6 -LA decomposition. Please note that the maximum temperature of 200 °C was reached after 45 minutes and was then kept constant until the end of the experiment.                 Table S4).

Figures
14 Figure S10. Modified Stern-Volmer plots for HSA fluorescence quenching by TBA-AlMo 6 -LA at A) 303 K, B) 308 K, and C) 310.5 K, respectively. The values for K A and n can be obtained from the intercept and slope of each plot, respectively (see Table S4).
14 Figure S10. Modified Stern-Volmer plots for HSA fluorescence quenching by TBA-AlMo 6 -LA at A) 303 K, B) 308 K, and C) 310.5 K, respectively. The values for K A and n can be obtained from the intercept and slope of each plot, respectively (see Table S4).  The interaction of the POM with HSA is described in more detail in the main text. C) The electrostatic surface potential of HSA. The electrostatic potential is measured in units of k B T/e, with a range as shown in the color bar. The environment of the binding site of AlMo 6 -LA (shown in ball-and-stick mode) is dominated by hydrophobic (white) and positively charged patches (blue) and therefore well suited to accommodate the POM with its negatively charged core and hydrophobic functionality. Please note that the second AlMo 6 -LA molecule, which is bound at a crystal contact, was ignored in this discussion because it most probably represents only a crystallographic artifact.
16 Figure S12. Further crystallographic details about the crystal structure of the HSA-Myr-AlMo 6 -LA. A) Electron density map (2mF o -DF c map contoured at 1.5σ) of the inorganic AlMo 6 core (Tris-lauroyl group is omitted for clarity). The disc-shaped electron density (shown as grey mesh) fits the POM molecule. B) The overall structure of the complex with the different domains of HSA being color-indicated. In addition, the location of Trp214 (shown as red sticks) and the fatty acid-binding sites (FA1-7) are indicated. TBA-AlMo 6 -LA binds to the interdomain cleft of HSA and exhibits contacts with domain IIIA, which is one of the two major drug binding sites (Sudlow site II); domain IIB; and a large loop that connects domain IA with domain IB. The alkyl chain of AlMo 6 -LA is sandwiched between helices of domain IB and IIIA, whereas the inorganic core interacts mainly with the before-described loop (including some hydrophobic interactions with one α-helix of domain IIIA). The interaction of the POM with HSA is described in more detail in the main text. C) The electrostatic surface potential of HSA. The electrostatic potential is measured in units of k B T/e, with a range as shown in the color bar. The environment of the binding site of AlMo 6 -LA (shown in ball-and-stick mode) is dominated by hydrophobic (white) and positively charged patches (blue) and therefore well suited to accommodate the POM with its negatively charged core and hydrophobic functionality. Please note that the second AlMo 6 -LA molecule, which is bound at a crystal contact, was ignored in this discussion because it most probably represents only a crystallographic artifact.
16 Figure S12. Further crystallographic details about the crystal structure of the HSA-Myr-AlMo 6 -LA. A) Electron density map (2mF o -DF c map contoured at 1.5σ) of the inorganic AlMo 6 core (Tris-lauroyl group is omitted for clarity). The disc-shaped electron density (shown as grey mesh) fits the POM molecule. B) The overall structure of the complex with the different domains of HSA being color-indicated. In addition, the location of Trp214 (shown as red sticks) and the fatty acid-binding sites (FA1-7) are indicated. TBA-AlMo 6 -LA binds to the interdomain cleft of HSA and exhibits contacts with domain IIIA, which is one of the two major drug binding sites (Sudlow site II); domain IIB; and a large loop that connects domain IA with domain IB. The alkyl chain of AlMo 6 -LA is sandwiched between helices of domain IB and IIIA, whereas the inorganic core interacts mainly with the before-described loop (including some hydrophobic interactions with one α-helix of domain IIIA). The interaction of the POM with HSA is described in more detail in the main text. C) The electrostatic surface potential of HSA. The electrostatic potential is measured in units of k B T/e, with a range as shown in the color bar. The environment of the binding site of AlMo 6 -LA (shown in ball-and-stick mode) is dominated by hydrophobic (white) and positively charged patches (blue) and therefore well suited to accommodate the POM with its negatively charged core and hydrophobic functionality. Please note that the second AlMo 6 -LA molecule, which is bound at a crystal contact, was ignored in this discussion because it most probably represents only a crystallographic artifact. Please note that all Anderson-type POMs in this table are double-side-functionalized structures (i.e., both sides of the POM core are functionalized).  The values of Table S4 were derived from Figure S10. Table S5 -α-helicity of free HSA and the HSA-TBA-AlMo 6 -LA-, HSA-AlMo 6 -, and HSA-LA-complexes, respectively.  Ramachandran outliers (%) 0.00
[b] R merge = Σ hkl Σ i |I i (hkl) -<I(hkl)>| / Σ hkl Σ i I i (hkl) [c] The mean intensity correlation coefficient of half-datasets. Values in parentheses are for the highest resolution shell.
[e] R free is calculated using a randomly chosen reference set of 5% of all the reflections collected.