Heterogeneously deacetylated chitosans possess an unexpected regular pattern favoring acetylation at every third position

doi:10.21203/rs.3.rs-3891399/v1

Download PDF

Article

Heterogeneously deacetylated chitosans possess an unexpected regular pattern favoring acetylation at every third position

https://doi.org/10.21203/rs.3.rs-3891399/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 06 Aug, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Chitosans are promising natural polymers with diverse industrial, medical and agricultural applications, but their properties and bioactivities depend on their structural characteristics, including their pattern of acetylation (PA). Chitosans produced by homogeneous deacetylation or chemical N-acetylation are assumed to have a random PA, but almost all commercially available chitosans are heterogeneously deacetylated, and whether they possess a random or block-wise PA has been the subject of a decades-long debate. Here we used a novel combination of analytical tools to address this unanswered question and found that both assumptions are wrong. Our combination of in vitro experiments and in silico modeling surprisingly revealed a more regular PA in heterogeneously deacetylated chitosans, with acetylated units overrepresented at every third position in the polymer chain. This unanticipated regular PA increases the elicitation activity of chitosans in plants compared to a random PA, and also generates different product profiles and distributions in enzymatic and acid hydrolysates. A regular PA may be beneficial for some applications but detrimental for others. Researchers in academia and in industries that use chitosans should therefore be aware that today’s commercial chitosans differ fundamentally from enzymatically produced natural chitosans and niche products produced by homogeneous deacetylation or N-acetylation.

Biological sciences/Biochemistry/Carbohydrates/Amino sugars

Biological sciences/Biochemistry/Carbohydrates/Polysaccharides

Biological sciences/Biological techniques/Analytical biochemistry/Liquid chromatography

Biological sciences/Biochemistry/Enzymes/Hydrolases

Biological sciences/Biotechnology/Biomaterials

Chitosans are versatile functional biopolymers with superior material properties and diverse biological activities, offering huge potential for the growing bioeconomy. These linear, polycationic polysaccharides consist of β-1,4-linked N-acetyl-d-glucosamine units (GlcNAc, A-units) and dglucosamine units (GlcN, D-units), thus differing in their degree of polymerization (DP), fraction of acetylation (FA), and pattern of acetylation (PA)¹. Chitosans with these differing structural characteristics possess diverse biological activities such as antimicrobial properties² and the ability to interact with cell surfaces and specific receptors^3–5. These activities are exploited in applications such as agriculture^6,7, the food and feed industry², water purification⁸, cosmetics⁹, and (veterinary) medicine, where they are used as wound dressings¹⁰ or to prepare nanoparticles¹¹ as vehicles for drug or vaccine delivery.

Chitosans are rare in nature, so commercial chitosans are derived from chitin, a homopolymer of β-1,4-linked GlcNAc units found abundantly in waste materials such as the exoskeletons of crustaceans and insects, the endoskeletons of cephalopods, and the cell walls of fungi¹². There are several ways to produce chitosans from chitin. First, chitin can be partially deacetylated in an alkaline solution, typically under heterogeneous deacetylation (HTDA) conditions (high concentration of NaOH, high temperature, and short reaction time)^13–15 but also under homogeneous deacetylation (HMDA) conditions (lower concentration of NaOH, moderate temperature, and longer reaction time)^16,17. Second, chitin can be completely deacetylated to polyglucosamine, and then treated by partial chemical N-acetylation (CNA) to produce chitosans¹⁸. It should be noted that chitin has two major crystalline polymorphic forms (α and β) that differ in the orientation of adjacent chitin sheets¹⁹. The more prominent α-chitin is found in crustacean shells, insect exoskeletons and fungal cell walls, whereas the endoskeleton of squids consists of β-chitin²⁰. Because α-chitin forms more intermolecular interactions than β-chitin, the latter swells faster in alkaline solutions and is therefore deacetylated more rapidly and efficiently¹⁴.

The widespread application of chitosans as multifunctional biopolymers has been limited by the poor reproducibility of observed effects, preventing the development of products with reliable performance. This mainly reflects the poor structural characterization of chitosans, given that the DP, FA and PA can all have a profound effect on the physicochemical properties²¹ and bioactivities¹ of these molecules. The full exploitation of chitosans in the bioeconomy depends on an in-depth understanding of their structure–function relationships. Standard methods are available to determine the DP²² and FA^23,24, and ‘second-generation chitosans’ with a defined DP and FA are now reaching the market. But analyzing the PA is more challenging. The current gold standard is the quantification of AA, DD, AD and DA diads by nuclear magnetic resonance (NMR) spectroscopy^13,25–27, which results in P_A values ranging from 0 (perfectly block-wise) through 1 (perfectly random) to 2 (perfectly alternating)^26,28. Less accurate methods include the analysis of DP distributions after chemical hydrolysis with HNO₂^29,30, or considerations based on solubility, crystallinity, enzymatic degradability, or behavior in solution or during dissolution^14,17,31,32. The research community agrees that the very rare commercial HMDA and CNA chitosans have a random PA^{13,17,25,26,30–32} but there is no consensus on the PA of HTDA chitosans (nearly all commercial chitosan products). There is some evidence for a block-wise PA^17,31–33, but most authors argue in favor of random PA^{13,25,26,29,34}, and some suggest that the deacetylation occurs first block-wise on the crystalline substrate and then continues in a random manner once the material becomes amorphous^14,30,35.

Enzymatic digestion followed by mass spectrometry (MS) is the most recent method for the fingerprinting of chitosans³⁶, and is based on the unique subsite preference of chitinosanase from Alternaria alternata (AaChio)³⁷. Whereas chitinases require an A-unit bound at subsite − 1 (GH18 chitinases) or subsite + 1 (GH19 chitinases), and most chitosanases require a D-unit at subsite − 2 or at subsites − 2 and − 1³⁸, AaChio has an absolute specificity for a D-unit at subsite − 2 and for an A-unit at subsite − 1. This enzyme therefore hydrolyzes chitosans precisely next to each transition from D- to A-blocks. Accordingly, AaChio-MS fingerprinting allows us to calculate the average A- and D-block sizes and block size distributions³⁶ (Supplementary Information, Eq. (S1) and Eq. (S2)).

Here we combined AaChio-MS fingerprinting³⁶ and size exclusion chromatography coupled to refractive index and MS detection (SEC-RI-MS) for product analysis^* to screen a large set of chitosans produced by different methods (HTDA, HMDA and CNA), aiming to answer the decades-old and still unresolved question: what is the PA of HTDA chitosans? Surprisingly, in vitro experiments combined with in silico modeling of chitosan populations and their enzymatic cleavage products clearly indicated that HTDA chitosans possess neither a random nor block-wise PA, but instead a more regular PA. We show that this influences the product profiles of hydrolytic enzymes as well as the bioactivity of chitosans in plants, and we propose a hypothesis to explain how this PA is formed during deacetylation. Given that commercially available chitosans are almost exclusively produced by HTDA, our findings are highly relevant when aiming to determine the structure–function relationships of these products to optimize their performance. Our data also suggest that commercial chitosans fundamentally differ in their structure from natural chitosans produced by the action of chitin deacetylases, which are highly unlikely to generate the same, regular PA.

Heterogeneously deacetylated chitosans differ from homogeneously deacetylated and chemically N-acetylated chitosans

Initially, we compared HTDA, HMDA and CNA chitosans with the same FA (0.32) by hydrolysis using AaChio followed by SEC-RI-MS analysis of the products. Unexpectedly, very clear differences were observed in the RI signals generated by the oligomeric products of the enzymatic digestion when separated by DP (Fig. 1a). The products of the HMDA and CNA chitosans showed a very similar uniform DP distribution, with the highest proportion of products covering the range DP 3–5, whereas the DP distribution of the HTDA chitosan products was strikingly different, with DP 3, 6, 9 and 12 detected in unexpectedly large amounts. To quantify the overrepresentation of these DP = 3n products, we introduced a parameter known as triad strength (Fig. 1b). This is calculated using Eq. (1), where I_{DP x} is the integrated RI signal of DP x (baseline from RI minimum between peaks for DP 6/7 and peaks for DP 2/3).

(1)\(triad strength= \frac{{I}_{DP 3}+{I}_{DP 6}}{{I}_{DP 2}+{I}_{DP 4}+{I}_{DP 5}}\)

The profiles of the AaChio products detected by MS also differed between the three types of chitosan (Supplementary Fig. S1a). Although absolute quantification is not possible due to differences in ionization efficiency between the oligomers, the data allow us to search for differences between samples, revealing that higher proportions of trimeric products (A₁D₂ and A₂D₁) were released from the HTDA chitosan compared to the others, consistent with the RI signals. We used these data to calculate the number average A- and D-block sizes (Supplementary Fig. S1b). Whereas the A-block sizes were comparable for all samples, the D-blocks were smaller for the HTDA chitosan. To rule out the unlikely possibility that these differences originate from different starting DP values, reflecting different weight average molecular weights (M_w) of the substrates (Supplementary Table S1), we compared AaChio products of HTDA and CNA samples with the same FA (0.17) and a similar M_w of ~ 110 kDa. The resulting RI chromatograms revealed the same trends in the FA = 0.17 samples (Supplementary Fig. S2) as seen for the FA = 0.32 samples (Fig. 1a). The corresponding triad strength values for the HTDA and CNA samples were 3.61 and 0.63, respectively.

Next, we checked whether these striking differences are indeed caused by the different production methods by analyzing a large set of HTDA, HMDA and CNA chitosans (Supplementary Table S1) spanning a range of FA values. Samples from different chitin sources were obtained from different producers, including chitosans that were deacetylated by freeze-pump out-thaw (FPT) cycles³⁵. This confirmed that the previously observed differences are valid without exception: the triad strength was < 1 for all CNA chitosans over the whole FA range of 0.07–0.57 (Fig. 2) whether they were produced at the University of Münster or University of Lyon, or by the commercial supplier Heppe Medical Chitosan (HMC). The HMDA chitosans called Viscosans, produced by Flexichem AB, also showed low triad strength values. In contrast, HTDA samples from two different commercial suppliers showed triad strength values up to more than 5. For sample series taken over the course of a HTDA reaction, it became apparent that the triad strength increased over the course of the reaction as the FA decreased, with the triad strengths of early samples being comparable to those of HMDA and CNA chitosans. In addition, we noticed differences in the deacetylation series produced from shrimp (α-chitin) and squid (β-chitin) starting materials. As expected from previous studies¹⁴, HTDA is more efficient for the less crystalline β-chitin (Supplementary Fig. S3), whereas the effect of high triad strengths at low FA values is more pronounced for chitosans derived from α-chitin. Samples prepared by the more efficient FPT process (HTDA)³⁵ of β-chitin behaved like other squid-derived HTDA chitosans. Interestingly, HMDA chitosans prepared from shrimp material by the Kjell Vårum group (Trondheim) show a behavior intermediate between shrimp-derived HTDA chitosans and CNA chitosans or HMDA chitosans from Flexichem AB (Viscosans).

Again, the differences between the HTDA, HMDA and CNA chitosans were also visible in the altered AaChio product profiles, which in turn lead to deviating number average A- and D-block sizes (Supplementary Fig. S4). For HMDA and CNA chitosans, the A-block size increases and D-block size decreases with increasing FA in a near linear manner over the entire FA range. HTDA chitosans show comparable block sizes for intermediate FA values > 0.3, but the A-block size is slightly higher and the D-block size slightly lower compared to CNA chitosans at lower FA values.

Overall, these data show striking differences between HTDA and HMDA/CNA chitosans, even for samples with comparable average FA and DP values. The strongest indicator is an overrepresentation of DP = 3n products, quantified by an increase in the triad strength, which was only observed for HTDA chitosans, and was stronger at lower FA values and more pronounced for chitosans derived from α-chitin.

Differences between chitosans are also apparent in chemical and other enzymatic hydrolysates

Next, we investigated the origin of the differences between HTDA and HMDA/CNA chitosans more closely by hydrolyzing the various FA = 0.32 samples using enzymes with different subsite preferences, as well as trifluoroacetic acid (TFA). The products were analyzed by SEC-RI-MS and the corresponding chromatograms are shown in Fig. 3. As discussed for the AaChio hydrolysates, HTDA chitosans were again cleaved into products where DP = 3n was overrepresented, but the strength of the effect varied between reactions. It was weakest for chitosanase CsnMN from Bacillus sp. MN^39–41 (Fig. 3e), which strongly prefers D-units at subsites − 2 and − 1, but more pronounced for the GH18 chitinases TvChi from Trichoderma virens⁴² (Fig. 3a) and human chitotriosidase (ChT)^43–45 (Fig. 3b), both of which have absolute specificity for A-units at subsite − 1. The effect was particularly strong for products of GH19 chitinases such as ChiG from Streptomyces coelicolor A3(2)⁴⁶ (Fig. 3c) and AtChi from Arabidopsis thaliana (TAIR: AT3G54420) (Fig. 3d), both of which have absolute specificity for Aunits at subsites − 2 and + 1. Interestingly, the preferred formation of DP = 3n products was visible not only in the enzymatic hydrolysates but also following the chemical hydrolysis of HTDA chitosans using TFA (Fig. 3f). Given that A-A bonds and A-D bonds are hydrolyzed three times faster than D-A and D-D bonds under acidic conditions⁴⁷, TFA hydrolysis corresponds to the behavior of an enzyme with a high preference for A-units at subsite − 1. The product profiles of all six reactions also differed between HTDA and HMDA/CNA chitosans (Supplementary Fig. S5). Furthermore, HTDA chitosan was hydrolyzed more efficiently by ChiG and less efficiently by ChT and TvChi, whereas the cleavage efficiency of all three chitosans was comparable for CsnMN and AtChi (Supplementary Table S2). Finally, the same trends were visible when HTDA and CNA chitosans of the same FA (0.17) and DP (M_w of ~ 110 kDa) were hydrolyzed (Supplementary Fig. S6).

In summary, the aberrant behavior of HTDA chitosans is visible in each hydrolysate, but the strength of the effect depends on the type of hydrolysis. This confirms that triad formation is not caused by the enzyme but is a property of the chitosan sample. Nevertheless, the sample must be hydrolyzed with certain specificity in terms of A-units and D-units to reveal the uniqueness of the HTDA samples, expressed by the irregular distribution of product DP values. Different average FA and/or DP values have been excluded as potential differences between the differently produced samples, and it would also be difficult to explain how these factors or their dispersity values (Đ_FA and Đ_DP) could lead to a higher proportion of trimers and hexamers in the hydrolysates. The only rational explanation is a PA in HTDA chitosans that differs from the random PA of HMDA or CNA chitosans and that possesses a certain regularity, which is reflected in the overrepresented products of the regular interval DP = 3n.

HTDA and HMDA/CNA chitosans and their hydrolysates show differences in bioactivity

To determine the significance of the more regular PA in HTDA chitosans in terms of bioactivity and potential applications, we tested the HTDA, HMDA and CNA chitosans (FA = 0.32) and their hydrolysates produced by chitosanase CsnMN^39–41 and chitinase TvChi⁴² for the ability to elicit an immune response in Arabidopsis thaliana seedlings. Compared to hydrolysates of HMDA/CNA chitosans, CsnMN cleaved the HTDA sample less efficiently and released small products of DP 2–6 that are more strongly deacetylated. TvChi was also less active on HTDA chitosan, and the respective DP 2–6 products appeared more highly acetylated than in the respective HMDA/CNA hydrolysates (Supplementary Fig. S7).

Oxidative burst assays were used to measure the amount of reactive oxygen species produced directly after treatment of A. thaliana seedlings with different concentrations of the polymers or hydrolysates as a central marker for an induced disease resistance reaction (Fig. 4). The HTDA polymer triggered the strongest oxidative burst and caused elicitation at low concentrations (Fig. 4a) whereas the CNA polymer had a weaker effect and higher concentrations were required for efficient elicitation. This tendency was even stronger for the HMDA polymer. Whereas the CsnMN hydrolysate of the HTDA sample was similar in activity to the undigested polymer, the CsnMN hydrolysis of the HMDA and CNA samples increased the oxidative burst (Fig. 4b). In contrast, enzymatic hydrolysis with TvChi reduced the elicitation activity of all three samples, especially the HTDA sample (Fig. 4c).

These results suggest that the more regular PA in HTDA polymers leads to a stronger immune response in plants compared to chitosan polymers with a random PA. For the latter, the cleavage of D-rich regions by CsnMN releases eliciting active products with A-rich centers, whereas the destruction of immunomodulatory A-rich parts by TvChi drastically reduces the elicitation activity*. Interestingly, hydrolysis of the HTDA sample using TvChi completely eliminated the initially very high bioactivity of this chitosan, although only limited hydrolysis occurred and a considerable amount of polymer remained. These differences in bioactivity indicate that the production of chitosans using different methods is interesting not only from an analytical perspective but also in the context of applications.

In silico models suggest A-units are present at every third position in HTDA chitosans

Our data strongly suggest that HTDA chitosans, contrary to current opinion, have neither a random nor a block-wise PA, but rather one with certain regularity, especially at lower FA values. But what PA would explain the preponderance of trimers, hexamers and nonamers? Deducing this from the products of CsnMN, TvChi, ChT, ChiG and AtChi is difficult because these enzymes show complex cleavage behaviors due to combinations of absolute specificities and preferences for A-units or Dunits at certain subsites. Therefore, we focused on the DP = 3n products of AaChio and chemical TFA hydrolysis. AaChio cleaves exclusively after the motif DA, whereas TFA hydrolysis preferentially occurs after an A-unit. Accordingly, the abundant DP = 3n products most often have an A-unit at their reducing end, corresponding to the patterns XXA, XXXXXA and XXXXXXXXA. For AaChio, the A-unit is preceded by a D-unit (XDA, XXXXDA and XXXXXXXDA). These products could in turn originate from a polymer in which A-units are overrepresented at every third position. We are not proposing that HTDA chitosans comprise the perfectly regular PA of (DDA)_DP/3 or even (XXA)_DP/3 but that the chitosans deviate from random and lean toward this more regular PA.

To model this, we used Python-based in silico methods to generate libraries of chitosan polymers with different average FA values and introduced different options for non-random PAs. The focus was on the overrepresentation of A-units at every third position to different extents, quantified by the parameter pattern strength ranging from 0 for a random PA to 1 for the PA closest to (DDA)_DP/3 for the corresponding FA. The cleavage of these chitosan libraries was then simulated at specific positions corresponding to the subsite preferences of different enzymes (e.g., after each DA-motif to simulate hydrolysis by AaChio)³⁷. The molar amounts of products of a certain DP released by AaChio from FA = 0.32 chitosans with different pattern strengths were then modeled in silico. For chitosan with a random PA, the products were evenly distributed with respect to DP (Fig. 5a). However, as soon as the overrepresentation of A-units at every third position was introduced (pattern strength 0.5), products of DP = 3n became more abundant compared to the products of neighboring DP values (Fig. 5b). This effect became more prevalent with increasing pattern strength up to the maximum value of 1 (Fig. 5c).

We also modeled how samples from later in the HTDA reaction might behave by limiting the deacetylation of units at every third position (which we named the pattern units) to FA = 0.75 while the overall FA decreased from 0.5 to 0.3. The extent of the overrepresentation of DP = 3n AaChio products increased with the decreasing average FA of the substrate, as observed in vitro (Supplementary Fig. S8). Similar to the simulated cleavage with AaChio, the in silico models for other hydrolysis reactions indicated that a PA deviating from random toward (DDA)_DP/3 (pattern strength > 0) will lead to an overrepresentation of DP = 3n products (Supplementary Fig. S9). It should be noted that for all in silico reactions, simplified absolute specificities of the enzyme families and the acid were assumed at certain subsites, and possible preferences at other subsites were ignored.

If the PA of HTDA chitosans is indeed this regular, why do previous studies conclude that HTDA chitosans have either a random^{13,25,26,29,34} or a block-wise PA^17,31–33, or combinations thereof^14,30,35? First, there are authors that justify their conclusions based on the different physicochemical properties and enzymatic degradability of the chitosans they tested^14,17,31,32. We argue that this deviating behavior may indeed indicate a non-random PA but not necessarily a block-wise PA as previously concluded. Second, DP distributions have previously been analyzed by SEC in the degradation products of differently produced chitosans following nitrous acid deamination^29,30, which results in cleavage exclusively after D-units⁴⁸. We modeled FA = 0.32 chitosan libraries with a random PA (pattern strength 0) as well as those with pattern strengths of 0.5 and 1, cleaved them in silico after D-units, and compared the DP distributions of the products (Supplementary Fig. S10). Just as previous studies reported, only small differences in the distributions for HTDA and HMDA chitosans and Bernoullian models for random PA, we found only small differences for pattern strengths of 0, 0.5 and even 1. Whereas the earlier studies claim that this indicates a random PA, we conclude that the method is not able to detect the more regular pattern as suggested here. Third, ¹H-NMR and ¹³C-NMR analysis of HTDA chitosans yielded diad frequencies comparable to those calculated for a random PA^13,25, or P_A values in the random-dominated range (0.5 < P_A < 1.5)²⁶. To investigate this, we modeled chitosan populations with different pattern strengths over the entire FA range from 0 to 1, and used them to calculate the P_A values as previously described^26,27 (Supplementary Fig. S11). Interestingly, the maximum P_A value (pattern strength 1, FA = 0.33), corresponding to the ideal (DDA)_DP/3 pattern, is only 1.5, so still within the previously defined random-dominated range²⁶. Furthermore, most commercial HTDA chitosans have FA values of ≤ 0.2 because customers favor low acetylated products. In combination with moderate pattern strengths, which we suspect to be more likely, the calculated P_A values deviate only negligibly from 1 (random PA). Therefore, diad analysis is appropriate to distinguish between alternating, random and block-wise PA^26,27, but cannot reveal longer repeated motifs, such as the pattern proposed here. Indeed, others have already suggested that diad analysis may report a random PA for certain block structures^25,36, which is why the NMR analysis of triad frequencies is also required (not to be confused with the terms triad strength and triad formation in our study). Again, earlier authors predicted random PAs for HTDA chitosans but they also reported that the method typically has errors of at least 15%²⁵. Our in silico models showed only slight deviations from random triad frequencies for small or intermediate pattern strengths (Supplementary Table S3), which fall within the typical error margins of the method. Accordingly, we conclude that both NMR-based diad and triad analysis are unable to detect the more regular PA in HTDA chitosans proposed herein.

We observed striking differences between the HTDA and HMDA/CNA chitosans at the structural and functional levels. Hydrolysates of all HTDA chitosans were enriched in DP = 3n products, although the extent of overrepresentation varied for different FA values and enzymes. In contrast, the hydrolysates of CNA and HMDA chitosans consistently showed a continuous distribution of DP values. Combining the in vitro experimental and in silico modeling data, we propose that HTDA chitosans exclusively feature a non-random but non-block-wise PA that deviates from random and tends toward the pattern (DDA)_DP/3, where A-units are overrepresented on every third unit. Our hypothesis is supported by earlier findings⁴⁹ revealing overrepresented DP = 3n products following the enzymatic digestion of HTDA chitosan, including the patterns (DXY)₁DAA, (DXY)₂DAA and (DXY)₃DAA (where X = A and Y = D or vice versa) for DP 6, 9 and 12, respectively. Our proposed (DDA)_DP/3 pattern, in combination with the enzyme’s subsite preference, would in fact lead to products such as (DDA)_nDAA or (DAD)_nDAA.

The origin of this special, regularly repeating PA in HTDA chitosans is unclear, although the degree of crystallinity in the original chitin material seems to play a role because the effect is more pronounced in chitosans derived from the more crystalline α-chitin rather than β-chitin. Moreover, homogeneous conditions during the production of HMDA chitosans from Flexichem AB (Viscosans) allow for the loss of crystallinity due to swelling, resulting in chitosans with a random PA. Accordingly, HMDA samples prepared using a less homogeneous process by the group of Kjell Vårum (Trondheim) show intermediate behavior between HTDA chitosans and Viscosans. If the regular PA is indeed formed during the HTDA reaction, there must be a molecular-level mechanism protecting every third position from efficient deacetylation. Decades ago, studies proposed that HTDA efficiently deacetylates the amorphous regions of chitin, while only slightly affecting the more crystalline portions^14,15,17. This reaction can be described by the shrinking core model^50,51, in which deacetylation occurs at the defined interface of the liquid alkali reagent and the solid chitin particle, leading to a shrinking core of still acetylated chitin that has yet to react. Simultaneously, the deacetylated chitosan forms a growing layer around the core and increasingly hinders inward hydroxide ion diffusion. The model explains why HTDA chitosans are suspected to have high FA dispersity (Đ_FA) values⁵² and why the efficiency of deacetylation levels off over the course of the reaction¹⁴, as we also observed (Supplementary Fig. S3). However, the shrinking core model does not explain how a regular PA could be formed. An alternative hypothesis for the cessation of deacetylation is the formation of a chitosan amide anion as part of the reaction mechanism⁵³. The anion is stabilized by the partially cationic acetamide group carbon of a neighboring chitosan chain, and its negative charge repels hydroxide ions that are necessary for subsequent deacetylation. This implies the formation of a regular secondary structure resulting from the intermolecular interactions of chitosan molecules during the HTDA reaction which, in turn, could account for the protection of specific positions from deacetylation and the formation of a regular PA. Another possible explanation is the formation of a stable interaction between a new D-unit and sodium acetate immediately after deacetylation, thus inhibiting the efficient deacetylation of surrounding units⁵⁴. In both cases, hydration (or more precisely the hydration status of hydroxide ions or chitin) may govern the reactivity. In conclusion, the HTDA reaction is complex and any or all of the reasons proposed above may explain the assumed high FA dispersity (Đ_FA), the cessation of the deacetylation reaction, and the formation of a more regular PA.

We note that the deviating PA in HTDA chitosans indicates neither inferior nor superior quality compared to HMDA and CNA chitosans. Rather, it is necessary to acknowledge these differences in PA, and that they lead to differences in chitosan bioactivity. Our assays with A. thaliana seedlings showed that the regular-PA HTDA polymer is a stronger elicitor than the random-PA HMDA and CNA chitosans of the same FA, suggesting that chitosans with a PA favoring (DDA)_DP/3 might be recognized particularly well by plant receptors⁵. Furthermore, all enzymatic hydrolysates of HTDA chitosans exhibited non-uniform DP distributions and product profiles deviating from those of HMDA and CNA chitosans. This is important because chitosans are usually processed by enzymes such as human ChT (in biomedical applications) or chitinases present in the soil microbiome or in plants (in agricultural applications). Our data suggest that even chitosans with the same average DP and FA and similar dispersities thereof can vary significantly in performance due to the different PAs resulting from different production methods. Thus, the special, regular PA we identified may be advantageous in some applications but a disadvantage in others.

Further research is needed to verify (or reject) the hypothesis of the (DDA)_DP/3 pattern formation and to find its molecular basis during the HTDA reaction. Analyzing more deacetylated HMDA or HTDA chitosans by AaChio hydrolysis and subsequent SEC-RI-MS could be useful, such as chitosans prepared using different process parameters, from different reaction time points, or from different chitinous starting materials. More importantly, it is necessary to examine the potential emergence of secondary structures during the HTDA reaction, which we suspect are responsible for the protection of every third unit from deacetylation.

In contrast to current opinion, we conclude that HTDA chitosans, including the majority of commercially available chitosans, have a non-random, non-block-wise PA and therefore differ considerably from random-PA chitosans produced by other methods. This is important because the PA influences the enzymatic processing, physicochemical properties^1,36,55, and bioactivities⁵⁶ of chitosans, thus affecting the performance of commercial products as well as the chitosans used for research purposes. Therefore, in addition to the structural parameters DP and FA (and ideally the corresponding dispersities, Đ_DP and Đ_FA), the process used to prepare chitosans (HTDA, HMDA or CNA) should be taken into account when investigating chitosan properties and bioactivities, and mentioned when reporting the obtained results.

Chitosan samples

The chitosan samples used in this study were either produced in-house or provided by Dr. Katja Richter (Heppe Medical Chitosan, Germany), Dr. Mats Andersson (Flexichem AB, Sweden), Prof. Kjell Vårum (Norwegian University of Science and Technology, Trondheim), or Gillet Chitosan SAS (France). Supplementary Table S1 provides detailed information about the origin, characteristics, production method, and chitin source. The FA was determined by enzymatic MS fingerprinting²⁴ (adapted method⁵⁷ for insoluble chitin samples), and the weight average and number average molecular weight (M_w and M_n, respectively) and dispersity (Đ_M) were analyzed by SEC-MALLS-RI²². The in-house HTDA samples were deacetylated at 90°C using 50% (w/v) NaOH for the indicated amount of time. Selected samples (series 2A) were then chemically partially depolymerized to obtain series 2B. FPT HTDA samples were prepared in-house as previously described³⁵, as were the CNA samples produced in Lyon and Münster²². The chitosans to be hydrolyzed with AaChio were dissolved overnight in ammonium acetate (200 mM, pH 4.2) whereas for other reactions, the chitosans were dissolved overnight in defined amounts of acetic acid corresponding to a molar ratio of acid to free chitosan amino groups of 1.2. Subsequently, dissolved chitosans were filtered with nylon centrifugal filters (pore size: 0.2 µm, VWR) to remove any insoluble particles.

Production of chitosan hydrolysates

The chitosans were hydrolyzed with AaChio³⁷ as previously described³⁶. Selected chitosans (Supplementary Table S1) with FA = 0.32 (HTDA, shrimp_100min; HMDA, Viscosan_DDA69; CNA, 134_0.29) or FA = 0.17 and M_w of ~ 110 kDa (HTDA, 80/20; CNA, 134_0.19) were also hydrolyzed at a concentration of 1 g/L with five other enzymes at 37°C: TvChi⁴², ChT^43–45, ChiG⁴⁶, AtChi (TAIR: AT3G54420), and CsnMN^39–41. Enzyme expression and purification as well as the precise reaction conditions for hydrolysis are provided in Supplementary Table S4. Chemical hydrolysis under acidic conditions followed the protocol for monosaccharide analysis⁵⁸. We mixed 20 µL of dissolved chitosans (1 g/L) with 1 mL TFA (2 M), autoclaved the reaction (45 min, 121°C) and air-dried the products at 40°C to remove the solvent. Twice, we added 500 µL of methanol, followed by air-drying at 40°C. The dried sample was dissolved in 30 µL of water for analysis.

SEC-RI-MS analysis

All hydrolysates were analyzed by SEC-RI-MS as described recently*. Briefly, 3 µg of each sample was injected and separated using an ACQUITY UPLC Protein BEH SEC column with a pore size of 125 Å (Waters Corporation), followed by RI detection using an ERC RefractoMax 520 (Thermo Fisher Scientific) and MS analysis using an amaZon speed ESI-MSⁿ (Bruker). The MS data were analyzed using Data Analysis v4.1 (Bruker) and an in-house Python script based on the module pymzML⁵⁹. Data Analysis v4.1 was also used to export RI data for further analysis using OriginPro 2023 (OriginLab). The calculation of block sizes was based on the AaChio products as previously described³⁶ (Supplementary Information, Eq. (S1) and Eq. (S2)).

Elicitation assay in A. thaliana

Three differently produced FA = 0.32 chitosans were hydrolyzed by adding 0.01 µg CsnMN or 0.1 µg TvChi per mg chitosan directly to the acidic aqueous chitosan solution and incubating at 37°C for three days or overnight, respectively. The hydrolysis was stopped by heating at 95°C for 5 min. The elicitation activity of the intact polymers and enzymatic hydrolysates was compared on A. thaliana seedlings as previously described⁵. Briefly, the samples were added to prepared seedlings and the response was measured by immediately quantifying the amount of produced H₂O₂ using luminol. Data represent the maxima of the burst curves.

In silico modeling

Python scripts were developed to simulate the behavior of different substrates and their hydrolysates in vitro. The foundation was the modeling of populations of chitosan molecules with varying characteristics (DP, average FA, and PA, including the strength of the (DDA)_DP/3 pattern). The modeled populations were used to count triad and diad frequencies and to calculate the P_A value from the latter, allowing for the simulation of diad-based PA analysis^13,25,26. They were also used for in silico hydrolysis by various enzymes with different specificities, HNO₂ (cleavage after D-units), or TFA (cleavage after A-units). The Python scripts are available as a user-friendly free web tool at https://lcp-simulator.anvil.app.

Data availability

The complete datasets generated and analyzed in this study are available upon request from the corresponding author.

Competing interests

The authors declare that they have no competing interests.

Author contributions

D.G. and S.T. prepared the samples, S.R. performed the elicitation experiments on A. thaliana and analyzed the results. M.J.H. performed all other experiments, developed and executed the Python scripts for in silico modeling, analyzed the data, and wrote the main manuscript. B.M.M. and S.C.-L. supervised the work. All authors discussed the results, developed hypotheses, and reviewed and revised the manuscript.

Acknowledgments

Margareta J. Hellmann would like to thank the German Academic Scholarship Foundation, which provided financial support for a doctoral scholarship. The authors acknowledge the financial support of the Open Access Publication Fund of the University of Münster and are grateful to Dr. Katja Richter (Heppe Medical Chitosan, Germany), Dr. Mats Andersson (Flexichem AB, Sweden) and Prof. Kjell Vårum (Norwegian University of Science and Technology, Trondheim) for the generous gift of chitosans. Special thanks to Prof. Martin Peter and Prof. Laurent David (University of Lyon, France) for very helpful critical discussions. We acknowledge Dr. Richard M. Twyman (Twyman Research Management) for editing the manuscript.

Wattjes, J. et al. Patterns matter part 1: Chitosan polymers with non-random patterns of acetylation. React. Funct. Polym. 151, 1–9; doi:10.1016/j.reactfunctpolym.2020.104583 (2020).
Raafat, D. & Sahl, H. G. Chitosan and its antimicrobial potential - A critical literature survey. Microb. Biotechnol. 2, 186–201; doi:10.1111/j.1751-7915.2008.00080.x (2009).
Goy, R. C., De Britto, D. & Assis, O. B. G. A review of the antimicrobial activity of chitosan. Polimeros 19, 241–247; doi:10.1590/S0104-14282009000300013 (2009).
Fuchs, K. et al. The fungal ligand chitin directly binds TLR2 and triggers inflammation dependent on oligomer size. EMBO Rep. 19, 1–14; doi:10.15252/embr.201846065 (2018).
Gubaeva, E. et al. ‘Slipped Sandwich’ Model for Chitin and Chitosan Perception in Arabidopsis. Mol. Plant-Microbe Interact. 31, 1145–1153; doi:10.1094/MPMI-04-18-0098-R (2018).
Malerba, M. & Cerana, R. Recent advances of chitosan applications in plants. Polymers (Basel). 10, 1–10; doi:10.3390/polym10020118 (2018).
Sharif, R. et al. The Multifunctional Role of Chitosan in Horticultural Crops; A Review. Molecules 23, 1–20; doi:10.3390/molecules23040872 (2018).
Yang, R., Li, H., Huang, M., Yang, H. & Li, A. A review on chitosan-based flocculants and their applications in water treatment. Water Res. 95, 59–89; doi:10.1016/j.watres.2016.02.068 (2016).
Rinaudo, M. Chitin and chitosan: Properties and applications. Prog. Polym. Sci. 31, 603–632; doi:10.1016/j.progpolymsci.2006.06.001 (2006).
Miguel, S. P., Moreira, A. F. & Correia, I. J. Chitosan based-asymmetric membranes for wound healing: A review. Int. J. Biol. Macromol. 127, 460–475; doi:10.1016/j.ijbiomac.2019.01.072 (2019).
Ali, A. & Ahmed, S. A review on chitosan and its nanocomposites in drug delivery. Int. J. Biol. Macromol. 109, 273–286; doi:10.1016/j.ijbiomac.2017.12.078 (2018).
Peniche, C., Argüelles-Monal, W. & Goycoolea, F. M. Chitin and Chitosan: Major Sources, Properties and Applications. in Monomers, Polymers and Composites from Renewable Resources (eds. Belgacem, M. N. & Gandini, A.) 517–542; doi:10.1016/B978-0-08-045316-3.00025-9 (Elsevier, 2008).
Vårum, K. M., Anthonsen, M. W., Grasdalen, H. & Smidsrød, O. Determination of the degree of N-acetylation and the distribution of N-acetyl groups in partially N-deacetylated chitins (chitosans) by high-field n.m.r. spectroscopy. Carbohydr. Res. 211, 17–23; doi:10.1016/0008-6215(91)84142-2 (1991).
Lamarque, G., Viton, C. & Domard, A. Comparative Study of the First Heterogeneous Deacetylation of α- and β-Chitins in a Multistep Process. Biomacromolecules 5, 992–1001; doi:10.1021/bm049780k (2004).
Lamarque, G., Viton, C. & Domard, A. Comparative Study of the Second and Third Heterogeneous Deacetylations of α- and β-Chitins in a Multistep Process. Biomacromolecules 5, 1899–1907; doi:10.1021/bm049780k (2004).
Sannan, T., Kurita, K. & Iwakura, Y. Studies on Chitin, 1: Solubility Change by Alkaline Treatment and Film Casting. Die Makromol. Chemie 176, 1191–1195; (1975).
Kurita, K., Sannan, T. & Iwakura, Y. Studies on chitin, 4. Evidence for formation of block and random copolymers of N-acetyl-D-glucosamine and D-glucosamine by hetero- and homogeneous hydrolyses. Die Makromol. Chemie 178, 3197–3202; doi:https://doi.org/10.1002/macp.1977.021781203 (1977).
Lamarque, G., Lucas, J.-M., Viton, C. & Domard, A. Physicochemical Behavior of Homogeneous Series of Acetylated Chitosans in Aqueous Solution: Role of Various Structural Parameters. Biomacromolecules 6, 131–142; doi:10.1021/bm0496357 (2005).
Jang, M.-K., Kong, B.-G., Jeong, Y.-I., Lee, C. H. & Nah, J.-W. Physicochemical characterization of α-chitin, β-chitin, and γ-chitin separated from natural resources. J. Polym. Sci. Part A Polym. Chem. 42, 3423–3432; (2004).
Kaya, M. et al. On chemistry of γ-chitin. Carbohydr. Polym. 176, 177–186; (2017).
Lamarque, G., Lucas, J. M., Viton, C. & Domard, A. Physicochemical behavior of homogeneous series of acetylated chitosans in aqueous solution: Role of various structural parameters. Biomacromolecules 6, 131–142; (2005).
Schatz, C., Viton, C., Delair, T., Pichot, C. & Domard, A. Typical Physicochemical Behaviors of Chitosan in Aqueous Solution. Biomacromolecules 4, 641–648; doi:10.1021/bm025724c (2003).
Hirai, A., Odani, H. & Nakajima, A. Determination of degree of deacetylation of chitosan by 1H NMR spectroscopy. Polym. Bull. 26, 87–94; doi:10.1007/BF00299352 (1991).
Wattjes, J., Niehues, A. & Moerschbacher, B. M. Robust enzymatic-mass spectrometric fingerprinting analysis of the fraction of acetylation of chitosans. Carbohydr. Polym. 231, 115684; doi:10.1016/j.carbpol.2019.115684 (2020).
Vårum, K. M., Anthonsen, M. W., Grasdalen, H. & Smidsrød, O. 13C-N.m.r. studies of the acetylation sequences in partially N-deacetylated chitins (chitosans). Carbohydr. Res. 217, 19–27; doi:10.1016/0008-6215(91)84113-S (1991).
Weinhold, M. X., Sauvageau, J. C. M., Kumirska, J. & Thöming, J. Studies on acetylation patterns of different chitosan preparations. Carbohydr. Polym. 78, 678–684; doi:10.1016/j.carbpol.2009.06.001 (2009).
Kumirska, J. et al. Determination of the pattern of acetylation of chitosan samples: Comparison of evaluation methods and some validation parameters. Int. J. Biol. Macromol. 45, 56–60; doi:10.1016/j.ijbiomac.2009.04.002 (2009).
Bovey, F. A. & Mirau, P. A. NMR of Polymers. (Academic Press, 1996).
Sashiwa, H., Saimoto, H., Shigemasa, Y., Ogawa, R. & Tokura, S. Distribution of the acetamide group in partially deacetylated chitins. Carbohydr. Polym. 16, 291–296; doi:10.1016/0144-8617(91)90114-R (1991).
Sashiwa, H., Saimoto, H., Shigemasa, Y. & Tokura, S. N-Acetyl group distribution in partially deacetylated chitins prepared under homogeneous conditions. Carbohydr. Res. 242, 167–172; doi:10.1016/0008-6215(93)80031-9 (1993).
Aiba, S. Studies on chitosan: 3. Evidence for the presence of random and block copolymer structures in partially N-acetylated chitosans. Int. J. Biol. Macromol. 13, 40–44; doi:10.1016/0141-8130(91)90008-I (1991).
Aiba, S. Studies on chitosan: 4. Lysozymic hydrolysis of partially N-acetylated chitosans. Int. J. Biol. Macromol. 14, 225–228; doi:10.1016/S0141-8130(05)80032-7 (1992).
Pacheco, N. et al. Structural Characterization of Chitin and Chitosan Obtained by Biological and Chemical Methods. Biomacromolecules 12, 3285–3290; doi:10.1021/bm200750t (2011).
Ottøy, M. H., Vårum, K. M. & Smidsrød, O. Compositional heterogeneity of heterogeneously deacetylated chitosans. Carbohydr. Polym. 29, 17–24; doi:10.1016/0144-8617(95)00154-9 (1996).
Lamarque, G., Cretenet, M., Viton, C. & Domard, A. New Route of Deacetylation of α- and β-Chitins by Means of Freeze−Pump Out−Thaw Cycles. Biomacromolecules 6, 1380–1388; doi:10.1021/bm049322b (2005).
Wattjes, J. et al. Enzymatic Production and Enzymatic-Mass Spectrometric Fingerprinting Analysis of Chitosan Polymers with Different Nonrandom Patterns of Acetylation. J. Am. Chem. Soc. 141, 3137–3145; doi:10.1021/jacs.8b12561 (2019).
Kohlhoff, M. et al. Chitinosanase: A fungal chitosan hydrolyzing enzyme with a new and unusually specific cleavage pattern. Carbohydr. Polym. 174, 1121–1128; doi:10.1016/j.carbpol.2017.07.001 (2017).
Davies, G. J., Wilson, K. S. & Henrissat, B. Nomenclature for sugar-binding subsites in glycosyl hydrolases. Biochem. J. 321, 557–559; doi:10.1042/bj3210557 (1997).
Weikert, T., Niehues, A., Cord-Landwehr, S., Hellmann, M. J. & Moerschbacher, B. M. Reassessment of chitosanase substrate specificities and classification. Nat. Commun. 8, 1698; doi:10.1038/s41467-017-01667-1 (2017).
Regel, E. K., Weikert, T., Niehues, A., Moerschbacher, B. M. & Singh, R. Protein‐engineering of chitosanase from Bacillus sp. MN to alter its substrate specificity. Biotechnol. Bioeng. 115, 863–873; doi:10.1002/bit.26533 (2018).
Gercke, D., Regel, E. K., Singh, R. & Moerschbacher, B. M. Rational protein design of Bacillus sp. MN chitosanase for altered substrate binding and production of specific chitosan oligomers. J. Biol. Eng. 13, 23; doi:10.1186/s13036-019-0152-9 (2019).
Bußwinkel, F., Goñi, O., Cord-Landwehr, S., O’Connell, S. & Moerschbacher, B. M. Endochitinase 1 (Tv-ECH1) from Trichoderma virens has high subsite specificities for acetylated units when acting on chitosans. Int. J. Biol. Macromol. 114, 453–461; doi:10.1016/j.ijbiomac.2018.03.070 (2018).
Eide, K. B. et al. Human Chitotriosidase-Catalyzed Hydrolysis of Chitosan. Biochemistry 51, 487–495; doi:10.1021/bi2015585 (2012).
Eide, K. B., Lindbom, A. R., Eijsink, V. G. H., Norberg, A. L. & Sørlie, M. Analysis of productive binding modes in the human chitotriosidase. FEBS Lett. 587, 3508–3513; doi:10.1016/j.febslet.2013.09.004 (2013).
Gorzelanny, C., Pöppelmann, B., Pappelbaum, K., Moerschbacher, B. M. & Schneider, S. W. Human macrophage activation triggered by chitotriosidase-mediated chitin and chitosan degradation. Biomaterials 31, 8556–8563; doi:10.1016/j.biomaterials.2010.07.100 (2010).
Heggset, E. B., Hoell, I. A., Kristoffersen, M., Eijsink, V. G. H. & Vårum, K. M. Degradation of chitosans with chitinase G from streptomyces coelicolor A3(2): Production of chito-oligosaccharides and insight into subsite specificities. Biomacromolecules 10, 892–899; doi:10.1021/bm801418p (2009).
Vårum, K. M., Ottøy, M. H. & Smidsrød, O. Acid hydrolysis of chitosans. Carbohydr. Polym. 46, 89–98; doi:10.1016/S0144-8617(00)00288-5 (2001).
Hussain, I., Singh, T. & Chittenden, C. Preparation of chitosan oligomers and characterization: their antifungal activities and decay resistance. Holzforschung 66, 119–125; doi:10.1515/HF.2011.130 (2012).
Bahrke, S. et al. Sequence analysis of chitooligosaccharides by matrix-assisted laser desorption ionization postsource decay mass spectrometry. Biomacromolecules 3, 696–704; doi:10.1021/bm020010n (2002).
Jiang, C. J. & Xu, M. Q. Kinetics of heterogeneous deacetylation of β-chitin. Chem. Eng. Technol. 29, 511–516; doi:10.1002/ceat.200500293 (2006).
de Souza, J. R. & Giudici, R. Effect of diffusional limitations on the kinetics of deacetylation of chitin/chitosan. Carbohydr. Polym. 254, 117278; doi:10.1016/j.carbpol.2020.117278 (2021).
Roberts, G. A. F. Chitosan production routes and their role in determining the structure and properties of the product. in Advances in Chitin Science Volume II (eds. Domard, A., Roberts, G. A. F. & Vårum, K. M.) 22–31; (Jaques André, 1997).
Novikov, V. Y., Derkach, S. R., Konovalova, I. N., Dolgopyatova, N. V. & Kuchina, Y. A. Mechanism of Heterogeneous Alkaline Deacetylation of Chitin: A Review. Polymers (Basel). 15, 1–23; doi:10.3390/polym15071729 (2023).
Lamarque, G., Chaussard, G. & Domard, A. Thermodynamic Aspects of the Heterogeneous Deacetylation of β-Chitin: Reaction Mechanisms. Biomacromolecules 8, 1942–1950; doi:10.1021/bm070021m (2007).
Sreekumar, S. et al. Biotechnologically produced chitosans with nonrandom acetylation patterns differ from conventional chitosans in properties and activities. Nat. Commun. 13, 7125; doi:10.1038/s41467-022-34483-3 (2022).
Basa, S. et al. The Pattern of Acetylation Defines the Priming Activity of Chitosan Tetramers. J. Am. Chem. Soc. 142, 1975–1986; (2020).
Urs, M. J., Moerschbacher, B. M. & Cord-Landwehr, S. Quantitative enzymatic-mass spectrometric analysis of the chitinous polymers in fungal cell walls. Carbohydr. Polym. 301, 120304; doi:10.1016/j.carbpol.2022.120304 (2023).
van Leeuwe, T. M. et al. A seven-membered cell wall related transglycosylase gene family in Aspergillus niger is relevant for cell wall integrity in cell wall mutants with reduced α-glucan or galactomannan. Cell Surf. 6, 100039; doi:10.1016/j.tcsw.2020.100039 (2020).
Kösters, M. et al. pymzML v2.0: introducing a highly compressed and seekable gzip format. Bioinformatics 34, 2513–2514; doi:10.1093/bioinformatics/bty046 (2018).

* reference missing, manuscript currently under review: Insights into chitosan-cleaving enzymes by simultaneous analysis of polymers and oligomers via UHPSEC-RI-MS; Margareta J. Hellmann, Bruno M. Moerschbacher, Stefan Cord-Landwehr

* reference missing, manuscript submitted in parallel to this manuscript

* reference missing, manuscript currently under review: LCP Simulator: An easy-to-use web tool to simulate binary linear copolymers and their behavior during analysis and enzymatic cleavage; Margareta J. Hellmann, Bruno M. Moerschbacher, Stefan Cord-Landwehr

There is NO Competing Interest.

Supplementaryinformation.pdf
Supplementary information

Download PDF

Journal Publication

published 06 Aug, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Heterogeneously deacetylated chitosans possess an unexpected regular pattern favoring acetylation at every third position

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Differences between chitosans are also apparent in chemical and other enzymatic hydrolysates

HTDA and HMDA/CNA chitosans and their hydrolysates show differences in bioactivity

Discussion

Materials and methods

Chitosan samples

Production of chitosan hydrolysates

SEC-RI-MS analysis

Declarations

Data availability

Competing interests

Author contributions

Acknowledgments

References

Footnotes

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1