Extraction, purification, and molecular weight characterization of AssP
Hot water was used to extract the polysaccharide from the stalk of the Chinese chive, and the crude polysaccharide was ethanol precipitated for its safety and simplicity for medicine and food products[17]. After pectin removal by CaCl2 titration, the polysaccharide was dialyzed to remove small molecules. The polysaccharide was further purified with gel filtration chromatograph, and each fraction of elution was detected. In our experimental setup, AssP eluted at around 270-330 mL with little impurities (Fig. 1, A).
Multiple methods were employed to measure the molecular weight of AssP, including GPC and MALDI-TOF/TOF mass spectrometry. Firstly, polysaccharide molecular standard of pullulan with a variety of molecular weight (6600, 21700, 113000, 348000, 805000 Da) was loaded to a GPC column and each elution time was obtained. The log value of their molecular weight was plotted against each corresponding elution time and a standard curve with R2 of 0.99 was obtained. The purified AssP was analyzed with identical procedures, and its elution peak corresponding to a time range of 9.1-9.8 min centered at 9.35 min (Fig. 1, B). This corresponded to a molecular weight range of 3121-434 Da centered around 1543 Da. To further identify the molecular weight of AssP, MALDI-TOF/TOF was employed to measure its accurate molecular weight. A cluster of peaks with 162 Da differences appeared on the mass spectrum in a range of 527.14-2147.38 Da, which is consistent with the GPC analysis (Fig. 1, C).
Structural characterization of AssP
To determine the monosaccharide composition of AssP, the polysaccharide was acid hydrolyzed before ion chromatography analysis. Monosaccharide standards of galactosamine, rhamnose, arabinose, galactose, glucose, mannose, fructose were analyzed in parallel to establish the curve for calculation of each mole fraction based on previous research[11]. The chromatogram of each above monosaccharide exhibited peak area that yielded a linear relationship with their concentrations that enabled us to calculate the corresponding mole fraction (Fig. 2, A, Additional File 1: Fig. S1, Table S1). Molar ratios of monosaccharides were calculated as 0.0264:2.46:3.71:3.35:1.00:9.93 of rhamnose: arabinose: galactose: glucose: mannose: fructose in the purified AssP (Fig. 2, B). It is worthy to mention that in the previous study[11], the ratio of 5.1:22.6:31.5:1 as arabinose: galactose: glucose: mannose was obtained for AssP. This is probably due to the incompetency of the 1-phenyl-3-methyl-5-pyrazolone (PMP)-derivation method they used for monosaccharide analysis, which could not detect the presence of fructose.
As shown in Fig. 3, A, the UV-Vis spectrum of gel filtration purified AssP yielded a significant absorption at 192 nm while no absorbance at 260 or 280 nm, indicating non-detectable nucleotide or protein in the purified AssP[18]. As shown in Fig. 3, B, functional groups of the AssP were characterized with FT-IR. The strong and broad bands at ~3392 cm-1 indicates the hydroxyl (O−H) stretching vibration of the polysaccharide chain. Bands at 2932 cm-1 were ascribed to the asymmetric C−H stretching vibration. Absorption at 1644 cm-1 was assigned to the carbonyl stretching of the polysaccharide. The peak at 1416 cm-1 corresponded to the C=O stretching and C−O bond from the carboxyl group and C−H bending vibrations. The band at 1024 cm-1 corresponded to the C−OH bonds, and the band at 928 cm-1 corresponded to the asymmetric stretching of pyranose ring in the polysaccharide chain.
Integrative analysis of a variety of 1D and 2D NMR spectrums lead to the identification of chemical shifts of sugar residues and their possible glycosidic linkage as listed in Table 1. HSQC spectrum was applied to identify the anomeric carbons, COSY and H2BC spectroscopy were applied to identify the 1H and 13C chemical shift of adjacent sugar residues, while HMBC and HSQC-TOCSY spectroscopy were applied to further identify the 1H and 13C correlation in between sugar residues or within one sugar residue. A total of 4 anomeric carbons could be identified, indicating the presence of sugar residues in 4 different chemical environments. Residue a exhibited one less carbon than residues b, c, and d. Thus, residue a was designated as the arabinose residue, and identification of other sugar residue b, c, and d were listed in Table 1. Since the chemical shifts between b and c are not distinguishable, they were designated as the same sugar residue. As illustrated in the monosaccharide component analysis and 1H spectroscopy (Fig. 4, A), sugar residues b and c should be galactose, and d should be glucose, respectively. In the one dimensional 13C, (Fig. 4, B), there was 3 clusters of quaternary carbon signals at around 103 ppm, and these signals were correlated with residues a, b, c, and d. Combined with all the information, it was designated as the characteristic signals of fructose. With information from all the spectroscopy, fructose e, f, and g were identified and listed in Table 1.
The hydrocarbon coupling constants of anomeric carbons in residue a, b, c, and d were all around 170 Hz, indicating the presence of α-glycosidic bonds[19]. In the HMBC spectroscopy (Table 1), correlation signals were found between a1 and f2, b1 and f2, d1 and e2, c1 and e2, suggesting linkage between a(1→2)f, b(1→2)f, d(1→2)e, and c(1→2)e. In the NOESY spectroscopy (data not shown), correlation between b5, c5, d5 and f4, g4 were observed. With all the information, the main backbone of the purified AssP could tentatively be α-L-Ara-(1→2)-β-D-Fru-(4→5)-α-D-G-(1→2)-β-D-Fru-(4→5)-α-D-G-(1→2)-β-D-Fru-(4→5)-α-D-G-(1→2)-β-D-Fru-(4→5)-α-D-G-(1→2)-β-D-Fru, among which G is Glc or Gal and the ratio between these two is approximately 1:1.
Tertiary structure and solution behavior of AssP
Polysaccharides containing triple-helical structures form a Cong-red-polysaccharide complex, and the maximum absorption wavelength would have red-shifted compared with Congo-red solution. Addition of strong alkali disrupts the hydrogen bonding thus the red-shift of the Congo red-polysaccharide complex is weakened[20]. The tertiary structure of purified AssP was examined with Cong red assay as shown in Fig. 5, A. The maximum absorption wavelength of the AssP-Conge red mixture did not yield any dramatic change (491 to 489 nm) as the concentration of NaOH increased, indicating the absence of triple-helical structure of AssP in solution.
The polysaccharide solution of each purification step was examined with DLS to reveal the solution behavior of AssP at each stage (Fig. 5, B). Water extraction of the Chinese chive yielded a mixture of multiple-size polysaccharide solution, with particle size ranging from 100 nm to 5000 nm. After precipitation with ethanol, the large proportion of polysaccharide with size ranging from 100 to 500 nm were homogenized to ~300 nm, while the largest particle at ~5000 nm did not change. Removal of pectin further homogenized the polysaccharide solution to a ~150 nm size particle solution as indicated by the narrower peak. Notably, gel filtration which further purified AssP rendered this polysaccharide to form a more homogenous, but larger size (with average size of ~600 nm) particle in solution. This was not mentioned in any studies previously, and it indicated a great potential of AssP being used as a natural encapsulating/delivering reagent in the future.
The AssP molecule in solution exhibited horseshoe-like configuration, and two AssP clamped each other at their grooves (Fig. 6, A). The third polysaccharide molecule ended clamping to the first one, with no specific binding site on neither of the AssP molecules. Up to ten AssP molecule were simulated to interact in solution. The fifth molecule joined the crowding by interacting with the third and the forth AssP, with no direct interaction with the very first one. Aggregation of further more AssP molecules exhibited a one-dimensional growth phenomenon where the binding of a later polysaccharide molecule always took place on the same site of the very first one. While stacking of these molecules yielded a certain curvature, nucleation at one side of the first AssP molecule would certainly form a spherical structure.
To visualize the resulting scenario of AssP aggregation, its solution state was simulated with sugar residues represented by coarse-grained bead to calculate their interactions (Additional File 1: Fig. S2, Table S2). This method emphasized on the solution distribution and particle formation (Fig. 6, B). At the very beginning of time (t=1step), the AssP molecules distributed homogenously in the aqueous solution. As time passed (t=500 steps), the polysaccharide molecules aggregated into irregularly sized particles. Large spherical particles began to show up at a time point (t=3000 steps), and the system tended to favor this type of solution behavior where eventually all the AssP molecules are distributed in the uniformly sized spherical particles (t=30000 steps). This is so far consistent with our DLS study (Fig. 5, B). However, it is likely a metastable state. At an approximate of infinite time point (t=50000 steps), the AssP spherical particles began to disassembly into differently sized particles, suggesting an eventually unstable assembly of this polysaccharide nano-particle.
The solution particle was also concentration dependent as the equilibrium state (t=50000 steps) of different concentrations of AssP were simulated (Fig. 6, C). At low concentrations, such as 1 mg/mL, the polysaccharide molecules form small spherical particles. As the concentration increased (10 mg/mL), the small particles approached each other and merged into larger particles. The spherical particles would merge into cylindrical micelle at higher AssP concentrations (20 mg/mL), and the diameter of the cylinder would increase as the concentration further increased (30 mg/mL). The diversity of AssP solution particle forms suggested a strong versatility in application.
Table 1 13C and 1H NMR assignment for AssP.
Residue
|
Nucleus
|
Chemical shift (ppm)
|
|
|
|
|
|
HMBC
|
JCH(Hz)
|
|
|
1
|
2
|
3
|
4
|
5
|
6
|
|
|
a
|
1H
|
5.34
|
3.44-3.44
|
3.68
|
3.36-3.40
|
3.74-3.77
|
|
|
|
|
13C
|
92.42
|
71.12
|
72.49
|
69.18
|
59.79-60.29
|
|
a1-f2
|
171.2
|
b
|
1H
|
5.33
|
3.44-3.45
|
3.67
|
3.41-3.47
|
3.83-3.87
|
3.70-3.74
|
|
|
|
13C
|
92.2
|
71.02
|
72.45
|
69.12
|
70.90-71.81
|
59.90-60.59
|
b1-f2
|
170.2
|
c
|
1H
|
5.32
|
3.46-3.47
|
3.66
|
3.36-3.40
|
3.84-3.87
|
3.74-3.77
|
|
|
|
13C
|
92.13
|
71.04
|
72.4
|
69.12
|
71.17-71.72
|
59.79-60.29
|
c1-e2
|
168.4
|
d
|
1H
|
5.31
|
3.47-3.48
|
3.65
|
3.41-3.47
|
3.83-3.87
|
3.70-3.74
|
|
|
|
13C
|
91.95
|
70.97
|
72.38
|
69.11
|
70.90-71.81
|
59.90-60.59
|
d1-e2
|
168.5
|
e
|
1H
|
3.58-3.61/3.65-3.69
|
|
4.08-4.11
|
3.98-4.02
|
3.76-3.81
|
3.6-3.75
|
|
|
|
13C
|
59.98-60.67
|
103.62
|
76.32-76.91
|
74.06-74.55
|
81.25
|
62.23
|
|
|
f
|
1H
|
3.60-3.64/3.71-3.74
|
|
4.18-4.20
|
3.95-3.99
|
3.8
|
3.6-3.75
|
|
|
|
13C
|
60.74-61.18
|
103.18
|
76.19-76.88
|
73.53-74.03
|
81.05
|
62.26
|
|
|
g
|
1H
|
3.78-3.81/3.62-3.65
|
|
4.12-4.16
|
4.03-4.07
|
3.75-3.80
|
3.6-3.75
|
|
|
|
13C
|
60.29-61.03
|
102.98
|
77.07-77.83
|
74.00-74.57
|
81.05
|
61.75-62.58
|
|
|