Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance

Abstract

Microalgae are photosynthetic organisms widely distributed in nature and serve as a sustainable source of bioproducts. Their carbohydrate components are also promising candidates for bioenergy production and bioremediation, but the structural characterization of these heterogeneous polymers in cells remains a formidable problem. Here we present a widely applicable protocol for identifying and quantifying the glycan content using magic-angle-spinning (MAS) solid-state NMR (ssNMR) spectroscopy, with validation from glycosyl linkage and composition analysis deduced from mass-spectrometry (MS). Two-dimensional ¹³C-¹³C correlation ssNMR spectra of a uniformly ¹³C-labeled green microalga Parachlorella beijerinckii reveal that starch is the most abundant polysaccharide in a naturally cellulose-deficient strain, and this polymer adopts a well-organized and highly rigid structure in the cell. Some xyloses are present in both the mobile and rigid domains of the cell wall, with their chemical shifts partially aligned with the flat-ribbon 2-fold xylan identified in plants. Surprisingly, most other carbohydrates are largely mobile, regardless of their distribution in glycolipids or cell walls. These structural insights correlate with the high digestibility of this cellulose-deficient strain, and the in-cell ssNMR methods will facilitate the investigations of other economically important algae species.

Figure 1

Microalgal carbohydrates found in starch grains, lipids, and cell walls. (a) Main components of microalgal cells. Carbohydrate-rich constituents (purple) include starch in grains, the galactolipids in the chloroplast and cytoplasm membranes, and polysaccharides or glycoproteins in the cell wall. Representative structures of algal carbohydrates are given in panels b–d. (b) Starch can be found in the amorphous or crystalline forms. The crystalline starch fraction is made of 6-fold double-stranded amylose helices, which packing and arrangement lead to two types of starch, i.e., the A- and B-types as revealed by X-ray crystallography.^, Amylose polysaccharide helices (green/red for carbon/oxygen) are less tightly packed and more hydrated (blue spheres) in the B-type, as compared to A-type starch. (c) The chloroplast inner membranes and other lipid bilayers contain mono- and digalactosyldiacylglycerol (MGDG, DGDG) and sulfoquinovosyldiacylglycerol (SQDG). (d) Cell walls in Chlorella and related algal species are rich in glucose (Glc), rhamnose (Rha), xylose (Xyl), ribose (Rib), galactose (Gal), mannose (Man), arabinose (Ara), and fucose (Fuc). Carbon numbers are given in full for Glc and only at key sites for other residues, with carbons 5/6 being the last carbon of a furanose/pyranose hexose unit.

Figure 2

Dynamic heterogeneity of microalgal molecules revealed by 1D ¹³C ssNMR spectra. The J-coupling-based refocused INEPT spectrum (magenta) only shows signals of highly mobile molecules. ¹³C DP spectra with a long recycle delay (30 s; black) enable quantitative detection while a short recycle delay (2 s, orange) detects mobile components. Arrows and stars are respectively used to highlight the carbon 1 sites of rigid starch and C1 of mobile carbohydrates. The difference spectrum (light blue) of the two DP parental spectra is highly similar to the CP spectrum (blue), selecting rigid molecules. All spectra were collected using intact P. beijerinckii cells on an 800 MHz NMR spectrometer, with 13.5 kHz MAS frequency and 83 kHz TPPM ¹H decoupling.

Figure 3

Complexity of carbohydrates structures resolved with 2D correlation spectra. (a) ¹³C refocused DP-based J-INADEQUATE spectrum of rehydrated uniformly ¹³C-labeled P. beijerinckii hydrated cells with a 2 s recycle delay selecting the mobile moieties of polysaccharides. The spectrum was acquired on an 800 MHz NMR spectrometer under 13.5 kHz MAS using 83 kHz TPPM ¹H decoupling. Each echo delay was set to 2.4 ms. Abbreviations are used for glycan resonance assignments and color-coding is primarily based on the symbol nomenclature for glycans (SNFG). The XZ^Y notation is used where X is the type of sugar, Y is the spin system number (for example, Ga¹, Ga², and Ga³ for 3 galactose units), and Z is the carbon number (for example, Ga6¹ is the carbon 6 of galactose 1 unit). Insets show the unique signals of arabinose carbon 1 and rhamnose carbon 6. (b) Representative structures of carbohydrate units with connectivity identified by mass spectrometry. The number in purple indicated for each unit corresponds to the relative EI detector response (percentage) obtained by GC-MS, which is a semiquantitative reporter as provided in Table S2.

Figure 4

Carbohydrate composition estimated from ssNMR and MS analyses of P. beijerinckii rehydrated cells. (a) Integration regions (green rectangles) of the 2D refocused DP-INADEQUATE spectrum used for glycan quantification (see Figure 3 for key experimental details and assignment). (b) Comparison of rehydrated P. beijerinckii CK-5 glycan composition derived from the 2D spectrum in panel (a) showing significant discrepancy with MS results. Data are summarized in Table S7. The difference is caused by the suppression of starch signals. (c) Good correlation between MS results and the recalibrated ssNMR quantification after correcting for the starch content. A correction factor of 3.5 obtained using 1D spectra (Figure S8) was applied on the starch fraction. (d) Cross sections extracted from the 2D refocused DP-INADEQUATE spectrum showing splittings for starch and most glycans. This observation is associated with ¹J_CC couplings. The starch slices shown here were deconvoluted (blue), with a good match between the simulated (magenta) and measured (black) spectra. Carbon 1 and terminal 5/6 each showed a single peak while other central carbons showed pairs of peaks.

Figure 5

2D ¹³C–¹³C correlation spectra for identifying intra- and inter-residue contacts in rehydrated P. beijerinckii cells. Sheared DP-INADEQUATE spectrum (gray) was overlapped with different CP-based spectra (RFDR, CORD, and PAR) that detect rigid molecules. (a) RFDR experiment with a short mixing time of 1.5 ms for selective detection of one-bond correlations. (b) 53 ms CORD detecting most of the intramolecular cross peaks. (c) 14 ms PAR showing long-range intra- and intermolecular cross peaks. (d) Subtraction of the CORD spectrum from the PAR spectrum resolving intermolecular contacts between starch and cell wall components. A scaling factor of 0.4 was applied to the CORD spectrum to cleanly remove all intramolecular cross peaks. All the CP-based spectra are dominated by starch, together with some of the galactose, xylose, and mannose units. One-bond correlations (dark purple), multibond intramolecular cross peaks (light purple), and inter-residue or intermolecular contacts (magenta) are marked and color-coded. (e) Numbers of peaks observed in different 2D spectra. The signals involved in one-bond correlations and their sugar types are shown for refocused DP- and CP-INADEQUATE, and RFDR spectra (top left). One-bond correlation peaks, multibond intramolecular cross peaks, and inter-residue or intermolecular contacts are shown for CORD, PAR, and their difference spectrum. All spectra were acquired with an 800 MHz NMR spectrometer under 13.5 kHz MAS frequency using 83 kHz TPPM ¹H decoupling.

Figure 6

2D ¹³C–¹³C correlation spectra spotlighting inter-residue and intermolecular contacts. (a) 14 ms PAR spectrum with long-range intra- and intermolecular cross peaks labeled. (b) Difference spectrum obtained from two parental spectra (PAR and CORD) detecting only intermolecular or inter-residue contacts. (c) Representation of identified inter-residue contacts, mainly from xylose (top), starch (middle), and galactose units (bottom). Details of the observed cross peaks are listed in Table S9. All spectra were acquired with an 800 MHz spectrometer using 13.5 kHz MAS frequency, with 83 kHz TPPM for ¹H decoupling.

Figure 7

Comparison of algal xylose with plant xylan. (a) Overlay of refocused INADEQUATE spectra measured with CP (orange) and DP (cyan) on P. beijerinckii whole cells, with an 800 MHz NMR spectrometer under 13.5 kHz MAS, with 83 kHz TPPM decoupling. Subtypes 4, 5, and 6 of P. beijerinckii xyloses show signals in both CP- and DP-based spectra, revealing their distribution in both rigid and mobile domains. (b) The signals of these algal xyloses are more similar to the 2-fold xylan (Xn^2f) observed in maize (gray) than to the 3-fold conformer (Xn^3f) reported recently. The reference spectrum of maize xylan was acquired with a 600 MHz spectrometer using a 14 kHz MAS frequency with 83 kHz TPPM decoupling. This maize spectrum was adapted with permission from Kang et al. Copyright (2019) Nature (https://creativecommons.org/licenses/by/4.0/).