Solid-state NMR analysis of unlabeled fungal cell walls from Aspergillus and Candida species

Graphical abstract


Introduction
The carbohydrate components in the fungal cell walls are promising therapeutic targets of novel antifungal drugs for combating lifethreatening infections by pathogenic fungi (Brown, 2012). Two families of compounds (echinocandins and terpenoids) have been developed to disrupt the biosynthesis of β-glucans in the cell wall, and another antifungal named nikkomycin is a potent inhibitor of chitin synthesis (Perlin, 2015;Denning, 2003;Jallow and Govender, 2021;Li et al., 2019;Perlin, 2020). Unfortunately, nikkomycin only has weak activity against most fungal species, and the β-glucan inhibitors often suffer from the compensatory paradoxical effect, through which the chitin level is massively elevated to compensate for the loss of β-glucans (Aruanno et al., 2019;Loiko and Wagener, 2017). Hence, an in-depth understanding of the biosynthesis and structure of fungal polysaccharides, as well as the structural dynamics of the assembled cell walls, may facilitate the discovery of more effective antifungal compounds from the structural perspective.
Solid-state NMR (ss NMR) spectroscopy is a non-destructive and high-resolution method for elucidating the structure of fungal biopolymers (Ghassemi, 2022). This method has been extensively employed to track melanization in Cryptococcus neoformans and other fungi (Zhong et al., 2008;Chatterjee et al., 2015;Chrissian, 2020). Studies have also been conducted on the analysis of Aspergillus fumigatus biofilm, with a focus on the quantification of carbon contributions in the extracellular matrix (Reichhardt, 2015;Reichhardt et al., 2019). Additionally, multidimensional ss NMR methods have been employed to investigate the nanoscale organization of A. fumigatus cell walls as well as its responses to stresses (Kang, 2018;Chakraborty, 2021). The samples being used in these studies are intact cells without chemical treatment, which allowed the investigation of biomolecules in a fully native physical state. This method was also combined with chromatography and mass spectrometry to understand the cell wall structure of Schizophyllum commune (Ehren, 2020).
Notably, in two recent ss NMR studies of A. fumigatus, a substantially revised structural scheme of the mycelial cell wall has been proposed (Kang, 2018;Chakraborty, 2021). The rigid and hydrophobic scaffold of the A. fumigatus cell wall is formed by tightly associated chitin and α-1,3glucan, which is dispersed in a mobile and hydrated matrix formed by branched β-1,3-glucans. Two other polysaccharides, including galactomannan and galactosaminogalactan, are found to mainly form the outermost dynamic layer, together with proteins on the cell surface. These structural aspects derived from ss NMR data provide a physical vision of polymer dynamics and packing, complementing the existing chemical analyses that reveal a chemical vision of the polymer crosslinking and extractability. Functional genomics approaches were also employed to generate mutant lines, each of which selectively remove an important carbohydrate component, to evaluate the functions of different polysaccharides and the remodeling of cell walls under internal stresses (Chakraborty, 2021). Statistical methods are also available for analyzing the polymorphic structure of polysaccharides in living fungal cells as demonstrated using chitin and chitosan (Fernando, 2021).
These ss NMR studies often require isotopic enrichment of the sample, e.g., by 13 C and 15 N, to allow 2D and 3D correlation experiments to be completed in a reasonable time frame, yet with adequate resolution. In fungal research, this has been achieved by germinating conidia into mycelia in minimal liquid media containing 13 C-glucose and 15 N sources. But there is a dire need for a method to validate the recent structural findings by examining samples prepared in solid media, a condition more widely used in microbiology research, in order to better integrate the structural concepts derived from ss NMR and biochemical approaches to promote sporulation (Chakraborty, 2021;Latgé and Wang, 2022). Also, the previous studies of A. fumigatus were primarily focused on the mycelium. It is questionable if the structural concepts could be applied to cells at different developmental stages, for example, the conidium. The inhaled conidia of A. fumigatus germinate into vegetative hyphae, invading the patient lungs during infection (Dagenais and Keller, 2009;van de Veerdonk et al., 2017). Even though it has been shown that the conidial and mycelial cell walls of A. fumigatus differ in their composition and organization , at this moment, we still lack a detailed molecular-level understanding of the conidial cell walls (Ibrahim-Granet, 2003).
Here we show that such challenges can be addressed by introducing natural-abundance (NA) magic-angle spinning dynamic nuclear polarization (MAS-DNP), which has been widely used in biopolymer and material research (Smith et al., 2019;Märker, 2017;Takahashi, 2012;Viger-Gravel, 2019;Perras, 2017), to better characterize fungal cell walls. This sensitivity-enhancing technique has made it possible to collect 2D correlation spectra using unlabeled samples. The NMR linewidths of most fungal carbohydrates are found to be substantially broadened at the cryogenic temperature of MAS-DNP due to their highly dynamic nature. A reserved carbohydrate core has been identified in A. fumigatus mycelia and conidia using unlabeled samples prepared in solid and liquid media, which could not be possible without the atomic resolution provided by the DNP-enabled 2D 13 C-13 C correlation spectra of these unlabeled materials. Further examination of polymer dynamics using conventional ss NMR at ambient temperature confirmed the recently proposed structural function of α-1,3-glucan in forming the stiff core of A. fumigatus cell walls in both mycelia and conidia. Finally, we extend this study to another prevalent fungal pathogen Candida albicans, demonstrating the applicability of these approaches in fungal research.

Preparation of fungal materials
Four unlabeled samples from A. fumigatus (RL 578) and C. albicans (SC5314) were grown using both liquid and solid media for MAS-DNP investigations. Unlabeled A. fumigatus materials were prepared in two ways using either solid or liquid media. The solid culture was prepared in YPD (Yeast-Extract Peptone Dextrose) agar. The liquid cultures were prepared in 100 mL of modified minimal media containing 10.0 g/L of glucose and 6.0 g/L of sodium nitrate. The pH of the media was adjusted to 6.5. Both liquid and solid cultures were incubated for 3 days at 30 • C (at 210 rpm). Similarly, both solid and liquid cultures were prepared for C. albicans without labeling. The solid culture was grown in YNB (Yeast Nitrogen Base) media with agar, 2 % of glucose, and 1 % of ammonium sulfate. The liquid culture was prepared using YNB, 2 % of glucose and 1 % of ammonium sulfate with pH adjusted to 5.8-6. The fungal material was harvested by centrifugation at 7000 g for 20 min. The fungal material was washed using 10 mM phosphate buffer saline (pH 7.4) to remove excess ions. Only the greenish center region of the solid culture was collected.
To validate the results obtained on unlabeled materials described above, we also prepared 13 C, 15 N-labeled samples using liquid media. The uniformly 13 C, 15 N-labeled A. fumigatus sample was prepared by adding 10.0 g/L of 13 C glucose and 6.0 g/L of 15 N-labeled sodium nitrate to the minimal liquid media . The uniformly 13 C, 15 Nlabeled C. albicans sample was prepared by adding 2 % of 13 C glucose and 1 % of 15 N-labeled ammonium sulfate into the YNB liquid media. Both strains were grown for 3 days at 30 • C.

Transmission electron microscopy and scanning electron microscopy
Transmission electron microscopy (TEM) was conducted using a JEOL JEM-1400 electron microscope. The sample was placed onto a glow discharged TEM grid for several minutes. It was stained using a mixture of 2 % uranyl acetate and lead citrate solution. The cell wall thickness was measured using ImageJ software (Schneider et al., 2012) after setting the scale in accordance with known bar scales on the cell images. Scanning electron microscopy (SEM) was conducted using an FEI Quanta 3D FEG field emission scanning electron microscope to examine the surface morphology of the cells. Briefly, cells were collected by filtration and fixed on a 0.4 μm pore polycarbonate filter in 2 % glutaraldehyde, 2 % formaldehyde, and 1 % OsO4. The sample was rinsed with distilled water, dehydrated with graded ethanol series, and dried with HMDS reagent. The cells were mounted to aluminum specimen stubs and coated with platinum in an EMS550X sputter coater for imaging.

Sample preparation for MAS-DNP measurements
Unlabeled fungal materials were mixed with the stock solution containing the biradicals needed for MAS-DNP. The stock solution contains 30 mM of AsymPol-POK biradicals (Mentink-Vigier, 2018;Harrabi, 2022) in 40 µL of d6-DMSO-D 2 O-H 2 O (1:8:1 vol%) that was used to avoid 13 C signal contribution from the solvents (e.g., from glycerol). The volume percentages of different solvents used here deviate from conventional recipes used for MAS-DNP of biomolecular samples. It is fully based on repeated optimizations of these fungal materials for the best sensitivity. To test the effect of the DNP juice, a different solvent of 13 C-depleted, d 8 -glycerol/D 2 O/H 2 O (6:3:1 vol%), as well as the AMUPol biradicals (Sauvée, 2013), were used for the samples prepared in solid media. The fungal samples were mildly ground using a set of mortar and pestle when being wetted by the cryoprotectant solution. This allows the radicals to penetrate and distribute in the porous cell wall, without perturbing the appearance of the fungal pellet and the molecular-level structure of molecules. Around 30 mg of fungal material was packed into 3.2 mm sapphire rotors.

MAS-DNP experiments
The MAS-DNP experiments were conducted on a 600 MHz (14.1 T)/ 395 GHz instrument using a 3.2-mm HCN DNP probe. The power of microwave irradiation was around 12 W. The temperature was ~ 100 K with microwave irradiation and decreased to 94 K when the microwave was turned off. The DNP buildup time constants were 2.6-s and 3.1-s for C. albicans samples prepared in solid and liquid media, respectively. The number was shortened to 1.3-s for the A. fumigatus solid sample. The MAS rate was 10.5 kHz for all DNP experiments unless mentioned otherwise. 1D 1 H-13 C cross-polarization (CP) experiments were measured using 1 H field of 50 kHz, providing a sideband match to the 13 C field of 39.5 kHz, and 1-ms contact time. 2D refocused dipolar (SPC5) (Hohwy et al., 1999) INADEQUATE spectra (Lesage et al., 1999) were collected under 10.5 kHz MAS. The field strength of 1 H decoupling during SPC5 blocks was 100 kHz. A Double-Quantum Filtered (DQF) 2D dipolar 13 C-13 C correlation NMR experiment (Wi et al., 2022) was carried out on A. fumigatus solid culture. The 1 H and 13 C 90-degree pulse lengths were 2.5-μs and 4-μs, respectively. In the indirect dimension, 200-300 increments were collected. In total, 64, 32, and 32 transients were added for signal-averaging purposes. The DARR mixing time was either 100-ms or 250-ms. The 1 H-13 C CP mixing was 0.5 ms. All the acquisition parameters are summarized in Supplementary Table S1.
Spectral deconvolution was performed on 1D 13 C CP MAS-DNP spectra of unlabeled liquid culture and solid culture of C. albicans to obtain the chemical shift, linewidth, and intensity of carbohydrate peaks. Spectral deconvolution was performed from 120 to 40 ppm for all the carbohydrate regions, using DmFit (Massiot, 2002). The parameters are provided in Supplementary Table S2.

Conventional solid-state NMR experiments at room temperature
The experiments were conducted on 800 MHz (18.4 Tesla) and 400 MHz (9.4 Tesla) Brucker spectrometers equipped with 3.2 mm and 4 mm HCN probes, respectively. All experiments were conducted under 13 kHz or 13.5 kHz MAS at 298 K temperature. Approximately 30 mg of sample was packed to 3.2 mm MAS rotors and around 110 mg of sample was packed in a 4 mm ZrO 2 rotor. The 13 C chemical shifts were externally referenced to tetramethylsilane (TMS) scale by calibrating the Cδ peak of the Met residue in the model tri-peptide N-formyl-Met-Leu-Phe-OH (MLF) to 14.0 ppm. The radiofrequency field strength was 83 kHz for 1 H decoupling and 62.5 kHz for 13 C hard pulse and 50 kHz for 1 H and 13 C CP spin lock. The acquisition parameters are tabulated in Supplementary Table S1.
To compare with the results of unlabeled samples, we measured 2D 13 C CP refocused J-INADEQUATE spectra using a uniformly 13 C-labeled A. fumigatus sample (liquid culture) on the 800 MHz spectrometer. We also measured a uniformly 13 C-labeled C. albicans sample (liquid culture), resulting in a 2D 13 C direct polarization (DP) refocused J-INAD-EQUATE spectrum measured on the 800 MHz spectrometer and a 13 C CP refocused dipolar (SPC5) INADEQUATE spectrum collected on the 400 MHz NMR. To investigate the polymer dynamics in the unlabeled fungal cell wall, 13 C-T 1 relaxation was measured. It was measured using CPbased T 1 pulse sequences (Torchia, 1978) with a z-filter varied from 0 to 5 s. The relative intensity of each data point (relative to the first data point) was plotted as a function of time. The curve was fit using a biexponential equation. For 13 C-T 1 relaxation measurements, the number of scans was between 1,024 and 4,096 for each data point of A. fumigatus samples prepared in solid and liquid media.

Morphological difference of fungal cells cultured in solid and liquid media
A. fumigatus cells have different morphologies when prepared in solid and liquid media (Fig. 1A) (Bernard and Latge, 2001;Fontaine, 2010;Latgé and Chamilos, 2019;Latgé, 2017;Kang et al., 2021). In a Petri dish containing YPD agar, A. fumigatus exhibited circular growth with greenish-blue conidia at the center and white mycelial threads at the edge ( Fig. 1A; top row). The surface morphology of the greenish-blue center region was examined using SEM images, which revealed a 2-3 μm diameter for the conidia of A. fumigatus (asexual spores) produced in conidiophores (fruiting body). Part of the outer cell wall of conidia should be covered with melanin, as indicated by the rough surface in the conidia observed in the zoomed-in regions of SEM images (Walker, 2010). The conidia from fungal plates were taken and inoculated into liquid minimal media and cultured at 30 • C for 3 days ( Fig. 1A; bottom row). The filamentous structures observed by SEM have confirmed that A. fumigatus mainly grew into hyphal form under this culture condition.
Evidently, the use of solid media has promoted asexual sporulation in A. fumigatus samples (Fig. 1A). A. fumigatus adapts to stressful environments during host interactions and acquires different morphotypes in their life cycle (Gow et al., 2002;Lee and Sheppard, 2016). The composition and organization of fungal cell walls are always changing in response to the morphotypes in the life cycle and growth conditions (Garcia-Rubio et al., 2020). The mycelium is the vegetative morphotype and the conidium is typically considered the infective morphotype (Latgé, 2017). The conidia of A. fumigatus disperse and colonize different habitats, for instance, the lung alveoli, germinating into hyphae and causing invasive infections (Bernard and Latge, 2001;Croft et al., 2016;Abad, 2010). Therefore, it is of high significance to elucidating the cell wall structure in conidia.
However, it is not trivial to convert the solid media (for example, yeast extracts) widely used in microbiology laboratories into uniformly 13 C, 15 N-labeled counterparts without worrying about the isotopedilution from unlabeled components. This barrier has hindered the use of high-resolution solid-state NMR spectroscopy for characterizing fungal conidia. At the same time, fungal materials cultured in minimal liquid media could not fully represent those prepared using complex media. This situation can be improved through the development of MAS-DNP techniques, as detailed in later sections.
After 3 days of incubation in YNB solid media, C. albicans produced cream-colored, dull smooth yeast-like colonies ( Fig. 1B; top row). SEM images revealed the oval shape of yeast-like C. albicans cells with diameters of 2-4 μm. In the liquid culture ( Fig. 1B; bottom row), C. albicans cells exist as a mixture of hyphae, pseudo hypha, and yeast forms, with the yeast form being the most prominent. The hyphae and germ tubes were present as minor components and hence were excluded from further consideration. Interpretation and conclusion in later sections are drawn by treating the yeast form as the overwhelmingly dominant form in the liquid cultures.
TEM images were used to quantify the distribution of cell wall thickness ( Fig. 1C and Supplementary Table S3) in A. fumigatus and C. albicans samples harvested from solid and liquid media. For A. fumigatus, the average thickness of the cell wall increased from 133 nm for the conidia (solid media) to 158 nm for the hypha (liquid culture). However, the change is much smaller between C. albicans materials prepared using solid and liquid conditions. Presently, it is not clear how the microscopic features of cellular morphology arise from the molecular-level organization of cell walls, warranting further investigations.  Table S4), making it possible to use short recycle delays, in turn resulting in the rapid acquisition of experiments. AsympolPOK biradicals were stable in fungal samples as confirmed by electron paramagnetic resonance (EPR) spectra ( Supplementary Fig. 1). The sensitivity was enhanced by 26 to 30 times for both A. fumigatus and C. albicans under microwave irradiation ( Fig. 2A, B), reducing the experimental time by 676-900 fold. Though the 1D 13 C peaks were very broad, some fine features of the peaks could still be discerned. The major signals are from chitin (Ch), β-1,3-glucan (B), and α-1,3-glucan (A) in A. fumigatus. For C. albicans, the prominent peaks emanated from chitin, β-1,3-glucan, and β-1,6-glucan (H). Determination of the linewidth and analysis of the nature of these peaks were aided by 2D 13 C correlation spectra, as described in detail later.

Sensitivity-enhancement of fungal materials by MAS-DNP
Notably, a record high sensitivity enhancement factor of 90-fold was achieved using an A. fumigatus sample doped with AMUPol using 13 Cdepleted, d 8 -glycerol/D 2 O/H 2 O solvent (Fig. 2C). The 13 C-depleted solvent is chosen to avoid the detection of glycerol signals that overlay with carbohydrate peaks. To our best knowledge, the 90-fold enhancement is the highest value reported for any cell wall system on a 600 MHz/395 GHz DNP, but this sample was not used for measuring 2D  experiments. The first reason is the prohibitively long DNP buildup time, 5.0 s for this sample instead of 1.3 s for the other A. fumigatus sample (Supplementary Table S4). A longer buildup time requires proportionally longer recycle delays, hence negating any sensitivity gain per unit time. Another consideration is the inhomogeneous hyperpolarization of this sample, with a preferentially higher enhancement for carbohydrate peaks relative to aromatic signals. It is a sign that the radicals were unevenly distributed across different polymer domains, and this spectrum may not accurately reflect the composition.

2D 13 C correlation spectra of unlabeled A. Fumigatus in solid and liquid media
The enhancement of NMR sensitivity gained by MAS-DNP allowed us to collect 2D 13 C-13 C correlation spectra using unlabeled A. fumigatus materials harvested from the solid media, which is a mixture of conidia and mycelia. In this study, the greenish conidia region was used for NMR measurements. Chemical analysis has reported that the core polysaccharides are conserved (though with variable composition) in the cell walls of hypha and conidium, but the conidium has additional layers of rodlets and melanin on the cell surface while the hypha is covered by galactosaminogalactan (Latgé and Chamilos, 2019). The chemical similarity of carbohydrate structure is informed by the similar spectral patterns of A. fumigatus materials cultured in liquid and solid media, providing an implication of the similarity in hyphae and conidia (Fig. 3A, B). Most of the carbohydrate signals observed in the 13 Clabeled samples of A. fumigatus liquid culture, which is hyphaedominant, showed corresponding signals in the natural-abundance 2D 13 C-13 C correlation MAS-DNP spectrum of the solid culture (Fig. 3B).
The chitin signals became stronger in the solid culture, as exemplified by the well-resolved peaks of chitin carbons 1 and 2 (Ch1 and Ch2) that resonate at 103 ppm and 55 ppm on the single-quantum (SQ) chemical shift dimension. This is an indication of a larger amount of chitin in the samples prepared using solid media. Chitin is among the most rigid molecules of A. fumigatus cell walls and its signals be preferentially detected by CP-based experiments at room-temperature. However, chitin signals are very weak in the CP refocused INADE-QUATE spectrum of liquid culture (Fig. 3A), consistent with its low intensities observed in two recent studies of A. fumigatus, which is likely due to the low content of this molecule in liquid cultures (Kang, 2018;Chakraborty, 2021). Signals were also observed for another two other major carbohydrates, α-1,3-glucans and β-1,3-glucans. The results suggest that at least the major constituents are shared between the hyphae and conidia. Many additional signals showed up in the DNP spectrum of the solid culture, likely from more mobile molecules, for example, galactomannan. Mobile carbohydrates are undetectable in the CP-based spectrum at room-temperature (Fig. 3A) but will become visible at the cryogenic temperature of MAS-DNP (Fig. 3B).
Unfortunately, the NMR linewidth, given by the full width at half maximum (FWHM), increased substantially at MAS-DNP conditions. The FWHM linewidths of resolved peaks in this 2D refocused INADE-QUATE spectrum were 2-3 ppm. This is considerably broader than the 0.5-0.9 ppm reported for samples measured at ambient temperature at a higher magnetic field (800 MHz) and is worse than the expectation for a spectrum collected on a 600 MHz ss NMR at room temperature. The spectral quality presented in Fig. 3B is much worse than that of unlabeled plant cell walls (Zhao et al., 2020;. This is due to the more dynamic nature of fungal cell walls compared to their counterparts in plants (Kang, 2018;Zhao et al., 2020). Crystalline components efficiently retain their sharp linewidth at 100 K during MAS-DNP measurements, which has been shown using the cellulose microfibrils in plant biomass (Zhao et al., 2020;. The major crystalline molecule in fungi is chitin, and it accounts for only 10-20 % of the dry mass of A. fumigatus cell wall (Lee and Sheppard, 2016). All other molecules, such as glucans, mannans, and exopolysaccharides, will suffer from the line-broadening effect due to their intrinsic disorder.
The most promising spectrum of unlabeled A. fumigatus was collected using a double-quantum filtered (DQF) 2D dipolar 13 C-13 C correlation scheme (Fig. 3C) (Wi et al., 2022). This experiment was finished in 30 hrs. The DQF-DARR spectrum of the solid A. fumigatus culture showed carbohydrate signals at 60-105 ppm, as well as unexpectedly strong signals of proteins, including both aliphatic carbons (0-70 ppm) and carbonyl groups (165-180 ppm). The quality of the DQF-DARR spectrum is manifestly superior to that of the CP refocused dipolar-INADEQUATE spectrum, considering both resolution and sensitivity as well as simultaneous detection of proteins and carbohydrates. In addition, the dipolar-INADEQUATE spectra collected at natural abundance often exhibit unmatched intensities for two carbons in a spin pair, such as the B1-B2 pair in Fig. 3B, which is not an issue in the DQF-DARR spectrum.
The DQF-DARR spectrum also exhibited satisfactory resolution, allowing us to observe signals from chitin and glucans (Fig. 4A). In addition to many one-bond correlations, the use of 100-250 ms DARR mixing enabled the detection of many medium-range cross peaks. An example is the C1-C3 cross peak of α-1,3-glucan (denoted as A1-3), which showed up at a unique position of (101 ppm, 84 ppm). In addition, the adequate sensitivity allowed us to assign the protein signals to different amino acid types by tracking the correlations among the Cα, Cβ, and CO (Fig. 4B). The chemical shifts are tabulated in Supplementary Table S5. Notably, some cross peaks between the aromatic carbons and the Cα and Cβ were also identified for aromatic amino acids (Fig. 4C). The performance of DQF-DARR is better than most other 2D 13 C-13 C correlation experiments, such as the refocused dipolar-INADEQUATE and CHHC that have been frequently employed to investigate unlabeled biomaterials (Viger-Gravel, 2019;Perras, 2017;Zhao et al., 2021). This DQF-DARR experiment might be critical to elucidating unresolved structural aspects in fungal cell walls.

Examination of biopolymer dynamics at room temperature
The dynamics of biopolymers in unlabeled A. fumigatus materials were probed using 13 C-T 1 relaxation, which was mapped using a series of 1D 13 C spectra with a gradually increasing z-period. This ss NMR approach has been widely used in the field of polymer research, where the samples are typically difficult to label. The information on polysaccharide dynamics afforded by this method perfectly complements the insight about polysaccharide structure obtained using naturalabundance MAS-DNP, as recently demonstrated on rice stems (Zhao et al., 2021).
Consistently observed in both the liquid and solid cultures, β-1,3-

Fig. 4. DNP-enabled 2D correlation spectra of unlabeled A. fumigatus solid culture (mostly conidia).
Selected regions of DQF 2D dipolar 13 C-13 C correlation spectra are shown for (A) carbohydrates, (B) protein aliphatic and carbonyl signals, and (C) aromatics. The CO presents the carbonyl group in proteins and the CX represents protein aliphatic carbons such as Cα and Cβ. CO-CX refers to the correlation between these carbon sites. The spectrum was acquired using the DQF-DARR sequence on a 600 MHz/395 GHz DNP. glucans have the fastest 13 C-T 1 relaxation (Fig. 5A), thus remaining the most mobile molecule in the cell walls, regardless of the fraction of hyphae and conidia. In contrast, α-1,3-glucans have the slowest relaxation, indicative of its immobility in A. fumigatus cell walls. The dynamics of chitin is between the αand β-1,3-glucans, being intermediately rigid. It should be noted that one carbon site of chitin, the C4 at 83 ppm, has exhibited very slow relaxation that is outside the range of the other carbons. This only happened to the sample prepared using solid media. We suspect that it is due to the insufficient resolution in 1D 13 C spectra, where the chitin C4 is influenced by the neighboring peak of α-1,3-glucan C3, which is just 1 ppm apart.
The bi-exponential feature became more pronounced in our unlabeled fungal materials (Fig. 5A) compared to the data collected on uniformly 13 C-labeled samples (Kang, 2018;Chakraborty, 2021). This is because the magnetization exchange between a pair of 13 C spins, mediated by 13 C -13 C spin interactions, gives rise to additional relaxation pathways, thus speeding up the relaxation process, which is no longer efficient at natural 13 C abundance. The inefficient exchange with mobile motifs also accounts for the substantially longer 13 C-T 1 time constants for the current unlabeled materials, than the uniformly 13 C-labeled samples used in recent studies (Kang, 2018;Chakraborty, 2021).
The dynamics can be better analyzed by comparing the 13 C-T 1 derived from the slow-relaxing component, corresponding to the less mobile domain of the biopolymer (Fig. 5B). The trends observed visually in Fig. 5A still hold. In each sample, the 13 C-T 1 time constants decreased successively for α-1,3-glucans, chitin, and then β-1,3-glucan. For example, the 13 C-T 1 time constants of these three molecules in the solid media were 13 ± 1 s, 11 ± 5 s, and 5.5 ± 0.3 s, respectively (Fig. 5B). On the other hand, 13 C-T 1 was similar in both the solid (5.5 ± 0.3 s) and liquid (5.0 ± 0.3 s) samples for β-1,3-glucans. This recurred for chitin molecules in different samples: the 13 C-T 1 remained similar in both solid and liquid cultures. In contrast, the 13 C-T 1 time constant of α-1,3-glucan decreased slightly from 17 ± 4 s in the liquid culture (hyphae) to 13 ± 1 s in the solid culture (mainly conidia). Therefore, α-1,3-glucans are slightly more dynamic in the conidia, though remaining the most rigid molecule across the cell wall.

Insight into the molecular organization of A. Fumigatus cell walls
It is generally accepted that the chitin molecule, due to its partial crystallinity, should be lending structural support to the fungal cell walls. This mechanical role is reminiscent of the function of cellulose in plant materials, consistent with its slow relaxation observed here. β-1,3glucans are the major cross-linking polysaccharides in A. fumigatus cell walls and are the key components for forming the chitin-β-1,3-glucanmannan core, a domain containing three covalently linked polysaccharides (Latgé, 2007). A branched analog, β-1,3/1,6-glucan, could also be introduced via the branching site of 3,6-linked glucose residue. With these considerations, it is not surprising that β-1,3-glucan stays as a relatively mobile polysaccharide to maintain its function of crosslinking and branching. The role and dynamics of β-1,3-glucan in the fungal cell wall are similar to those of hemicellulose and pectin (together, named matrix polysaccharides) in the primary plant cell walls (Zhao et al., 2020).
For decades, α-1,3-glucan has remained a mysterious molecule in A. fumigatus cell walls. It is largely extractable by hot alkali, thus lacking covalent bonds to other components, which has led to the assumption that α-1,3-glucan is less important to the cell wall organization (Bernard and Latge, 2001;Latgé, 2007). This chemical view has been revamped recently, where a large number of cross peaks (physical contact on the sub-nanometer scale) were identified between chitin and α-1,3-glucan.
The ss NMR data support a model in which chitin and the majority of α-1,3-glucan are tightly associated to form rigid aggregates that exclude water molecules to a large extent (Kang, 2018;Chakraborty, 2021). Though this paradigm has come under scrutiny (Kang, 2018;Chakraborty, 2021;Latgé and Wang, 2022), the extremely slow 13 C-T 1 relaxation of α-1,3-glucan observed in both liquid and solid cultures, with the latter being more relevant to the conditions used in most microbiology studies (Kang et al., 2021;Gastebois et al., 2009;Ashton and Dyer, 2019), confirms the stiffness of most α-1,3-glucans in A. fumigatus cell walls.

An exploratory investigation of unlabeled C. albicans samples
We further applied the natural-abundance MAS-DNP method to examine unlabeled C. albicans samples. Satisfactory enhancement factors of 26-and 30-fold were obtained from two samples harvested from liquid and solid media, respectively (spectra not shown). Based on the sensitivity, we managed to collect nearly noiseless 1D 13 C spectra (Fig. 6A), which showed high similarity between the solid and liquid cultures as evidenced by their consistent spectral patterns. This leads to the inference that the growing conditions have relatively minor effect on the composition of cell walls. Spectral deconvolution allowed us to disentangle the underlying carbohydrate resonances (Fig. 6A), and this process is assisted by the resolvable carbon sites in the high-resolution 2D 13 C-13 C correlation spectra enabled by MAS-DNP (Fig. 6B). To facilitate the assignment, the 2D 13 C correlation DNP spectrum of unlabeled C. albicans was overlaid with a high-resolution spectrum collected at ambient temperature on a 13 C-labeled sample (Fig. 6B). In the liquid culture, this comparison allowed us to resolve the signals from β-1,3-glucans, β-1,6-glucans, chitin, as well as some mannose units, which likely originate from peptidomannans in C. albicans cell walls (Hall and Gow, 2013). Consistent with the 1D spectra, the 2D spectra also showed a high level of consistency between the cell walls from solid and liquid cultures. Though the spectral resolution is still limited, these observations constitute an early insight into the structural similarity of the core polysaccharides in the yeast form and hyphae of C. albicans, and demonstrate the feasibility of using MAS-DNP to investigate different fungal strains. Notably, the C. albicans cells mainly exist in the unicellular form (yeast form) observed here. The fungal cell walls, especially those of C. albicans, are substantially more mobile than the plant materials and are considered unfavorable for MAS-DNP. The success of natural-abundance DNP thus opens the frontier to interrogate other cellular systems with similar dynamical characteristics, such as microalgae, bacteria, and human and animal cells (Takahashi, 2013;Overall and Barnes, 2021;Albert et al., 2018;Romaniuk and Cegelski, 2015;Poulhazan, 2021;Poulhazan et al., 2018;Arnold et al., 2018;Ghosh et al., 2021), without isotopic enrichment.

Conclusions
We have shown that the MAS-DNP enabled 2D 13 C-13 C correlation experiments of unlabeled fungal materials provide a way for addressing important structural questions that would remain unanswered otherwise. The fungal cell wall may however not be deemed to be the most ideal system for MAS-DNP due to the highly dynamic nature of most fungal carbohydrates (in contrast to plant cell walls) and the significant line-broadening at cryogenic temperature. Despite the difficulty, this strategy has led us to show the similarity of the major polysaccharides in unlabeled fungal materials prepared from solid and liquid media, which could not be possible without MAS-DNP. The carbohydrate fingerprints are also found to be consistent in the conidia and hyphae of A. fumigatus. Still, more experiments, especially those designed to measure intermolecular packing, are needed for further assessing the difference in the nanoscale organization. Such development could pave the way for investigating fungal materials that are difficult to label or replicate in the lab, such as disease-relevant fungal isolates and those requiring coculture with human and animal cells.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 6. DNP measurement of C. albicans samples. (A)
Spectral deconvolution of 1D 13 C CP spectra measured on C. albicans samples prepared in liquid cultures (left) and solid media (right). The spectra are measured on a 600 MHz/395 GHz DNP system. The simulated spectra (red) match well with the experimentally measured spectra (blue). Underneath are many individual lines that add up to the simulated spectra. (B) 2D 13 C refocused INADEQUATE spectra of C. albicans. The left panel shows the CP refocused dipolar-INADEQUATE spectrum of unlabeled liquid culture measured using 600 MHz/395 GHz DNP (black grey) and the right panel shows the same type of spectrum collected on unlabeled C. albicans materials grown in solid media. For each panel, the DNP spectrum is overlaid with two highresolution 2D spectra measured on 13 C-labeled liquid culture at room temperature (RT), including a DP refocused J-INADEQUATE spectrum measured on an 800 MHz NMR (blue) and a CP refocused dipolar-INADEQUATE spectrum measured on a 400 MHz ss NMR (yellow). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)