Degradation of chitosans with chitinase B from Serratia marcescens Production of chito-oligosaccharides and insight into enzyme processivity Audun Sørbotten 1 , Svein J. Horn 2 , Vincent G. H. Eijsink 2 and Kjell M. Va ˚ rum 1 1 Norwegian Biopolymer Laboratory (NOBIPOL), Department of Biotechnology, Norwegian University of Science and Technology, Trondheim, Norway 2 Department of Chemistry, Biotechnology and Food Science, Agricultural University of Norway, A ˚ s, Norway Chitosans are linear, cationic polysaccharides composed of (1 fi 4)-linked units of 2-amino-2-deoxy-b-d-gluco- pyranose (GlcN, D-unit) which may be N-acetylated to varying extents. Chitosans are soluble in acidic solution, whereas chitin, a structural polysaccharide occurring mainly in the exo-skeleton of arthropods, is insoluble [1]. Chitin is composed of (1 fi 4)-linked units of 2-ace- tamido-2-deoxy-b-d-glucopyranose (GlcNAc, A-unit). Chitin shares many similarities with cellulose, e.g. the conformation of the monomers, di-equatorial glycosidic Keywords chitinase B; chitosan degradation; oligosaccharides; processive enzymes; subsite analysis Correspondence K. M. Va ˚ rum, Norwegian Biopolymer Laboratory (NOBIPOL), Department of Biotechnology, Norwegian University of Science and Technology, 7491 Trondheim, Norway Fax: +47 73591283 Tel: +47 73593324 E-mail: kjell.morten.vaarum@biotech.ntnu.no (Received 12 October 2004, accepted 18 November 2004) doi:10.1111/j.1742-4658.2004.04495.x Family 18 chitinases such as chitinase B (ChiB) from Serratia marcescens catalyze glycoside hydrolysis via a mechanism involving the N-acetyl group of the sugar bound to the )1 subsite. We have studied the degradation of the soluble heteropolymer chitosan, to obtain further insight into catalysis in ChiB and to experimentally assess the proposed processive action of this enzyme. Degradation of chitosans with varying degrees of acetylation was monitored by following the size-distribution of oligomers, and oligomers were isolated and partly sequenced using 1 H-NMR spectroscopy. Degrada- tion of a chitosan with 65% acetylated units showed that ChiB is an exo- enzyme which degrades the polymer chains from their nonreducing ends. The degradation showed biphasic kinetics: the faster phase is dominated by cleavage on the reducing side of two acetylated units (occupying subsites )2 and )1), while the slower kinetic phase reflects cleavage on the reducing side of a deacetylated and an acetylated unit (bound to subsites )2 and )1, respectively). The enzyme did not show preferences with respect to acetyla- tion of the sugar bound in the +1 subsite. Thus, the preference for an acetylated unit is absolute in the )1 subsite, whereas substrate specificity is less stringent in the )2 and +1 subsites. Consequently, even chitosans with low degrees of acetylation could be degraded by ChiB, permitting the pro- duction of mixtures of oligosaccharides with different size distributions and chemical composition. Initially, the degradation of the 65% acetylated chitosan almost exclusively yielded oligomers with even-numbered chain lengths. This provides experimental evidence for a processive mode of action, moving the sugar chain two residues at a time. The results show that nonproductive binding events are not necessarily followed by substrate release but rather by consecutive relocations of the sugar chain. Abbreviations ChiB, chitinase B; DP, degree of polymerization; D-units, GlcN-units; A-units, GlcNAc-units; a, degree of scission; F A , fraction of acetylated units. 538 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS linkages, crystalline structure, and lack of solubility in aqueous solvents. The detailed chemical composition of chitosans will vary, depending on the fraction of acetylated units (F A ) and their synthesis or preparation. This variation affects many properties of chitosan, such as solubility as a function of pH [2], binding to lysozyme [3], sus- ceptibility to degradation by lysozyme [4,5], as well as functional properties in drug delivery [6] and gene delivery systems [7]. Previous studies have shown that chitosans obtained by homogeneous de-N-acetylation of chitin, such as those used in the present study, have a random distribution of N-acetyl groups [8–10]. Chitinases and chitosanases are capable of convert- ing chitin and chitosans to low molecular mass prod- ucts (oligosaccharides) by hydrolyzing the b(1 fi 4) glycosidic linkages between the sugar units. There are several families of chitinases (glycoside hydrolase fam- ilies 18 and 19) and chitosanases (glycoside hydrolase families 46, 75 and 80) [11]. These enzymes differ with respect to their preference for A- and D-units on each side of the scissile glycosidic bond and, consequently, differ with respect to their activities towards different types of chitosan. Like many other chitinolytic bac- teria, the Gram-negative soil bacterium Serratia mar- cescens produces multiple chitinases (ChiA, ChiB, ChiC1 and ChiC2) [12–14]. These enzymes all belong to family 18 chitinases, which have a characteristic cat- alytic mechanism which depends on participation of the N-acetyl group of the sugar unit bound to the )1 subsite [15–19]. This substrate-assisted catalytic mech- anism implies that family 18 chitinases are expected to have an absolute preference for A-units in the )1 sub- site and that the presentation of a deacetylated sugar (D-unit) to the )1 subsite would represent nonproduc- tive binding. Chitinase B (ChiB; EC 3.2.1.14) from S. marcescens is a two-domain family 18 chitinase [20]. The active site of ChiB has defined subsites from )3 to +3 [16,20], and the substrate-binding cleft is tunnel-like, permitting tight interaction with the polymeric substrate. The aglycon- side of the substrate-binding groove is extended by the aromatic surface of a putative chitin-binding domain. The crystal structure of ChiB further suggests that the active site cleft is partially blocked at the )3 subsite [20]. Together with studies showing that ChiB converts chitin primarily to dimers [14,21], the structure of ChiB sug- gests that the enzyme exerts exo-activity and degrades chitin chains from their nonreducing ends [16,20]. It is generally accepted that enzymes that degrade carbohydrate polymers can do so by three principally different mechanisms, as first described by Robyt and French [22,23]: (a) A multiple-chain mechanism, where the enzyme-substrate complex dissociates after each reaction. (b) A single-chain mechanism, where the enzyme remains associated with the substrate until every cleavable linkage in the chain has been hydro- lyzed. (c) A multiple attack mechanism, where a given average number of attacks are performed after the enzyme-substrate complex has been formed. The latter two mechanisms are often referred to as ‘processive’. Carbohydrate degrading enzymes with tunnel-like act- ive site clefts are often suggested to act processively [24]. For example, cellobiohydrolases are thought to degrade cellulose according to a multiple attack mech- anism, where dimers are cleaved off from the end of the polysaccharide chain [24–27]. These cellulose- degrading enzymes convert cellulose to cellobiose, along with small amounts of monomers and trimers, which are thought to be the result of the initial bind- ing ⁄ cleaving event. Whereas plausible models exist for cellulase action, mainly based on the results of site- directed mutagenesis and structural studies, less is known about processivity in chitinases. The only avail- able information comes from Uchiyama et al. [28], who used electron microscopy to study the degradation of b-chitin fibers by ChiA from S. marcescens and who concluded that their results were compatible with a processive mode of action. One goal of the present study was to exploit the unique heteropolymeric nature of chitosan to study the mechanism of action of ChiB. A second goal of this study was to explore possibilities of producing chito- oligosaccharide mixtures by degrading chitosans with ChiB. This latter goal is of interest because chito- oligosaccharides have a number of potential applica- tions [29–32]. To obtain these goals, we have analyzed the degradation of well-characterized, fully water- soluble, high molecular mass chitosans with ChiB, using a newly developed chromatographic method for oligomer separation and previously established NMR- based methods [5,9,10,33] for analysis of oligomer com- position. For example, we studied the time course of depolymerization of a highly acetylated but still water- soluble chitosan of high molecular mass, and the result- ing oligosaccharides were characterized with respect to their chain length and chemical composition. In addi- tion, high molecular mass chitosans of varying F A were degraded, and the resulting oligosaccharides were char- acterized in the same way. The results show that ChiB acts as an exo-chitinase ⁄ -chitosanase with specific requirements towards acetylated ⁄ deacetylated sugar units in subsites )2 and )1, which is reflected in the formation of oligomers with widely varying size distri- butions and chemical compositions, depending on the type of chitosan used. The results also provide unequi- A. Sørbotten et al. Degradation of chitosans with ChiB FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 539 vocal experimental evidence for a processive mode of action in ChiB. Results Size-exclusion chromatography of chito- oligosaccharides The size-exclusion chromatographic system was exten- sively calibrated using fully acetylated and fully de-N-acetylated oligomer standards. Figure 1 shows the degree of polymerization as a function of the dis- tribution coefficient ½K av ¼ðV e À V 0 Þ=ðV t À V 0 Þ where V e is the elution volume of the actual oligo- mer, and V 0 and V t is the void volume and the total volume of the column, respectively. Although the two homologous oligomer-series do not elute at exactly the same K av values, the calibration lines are close and nearly parallel. Extrapolation of the calibration lines to K av ¼ 0 revealed that the corresponding degree of polymerization (DP)-value was 40. The cal- ibration lines of the acetylated- and the de-N-acetylat- ed oligomers did not intercept (an interception of the calibration lines would mean that a fully deacetylated oligomer would co-elute with an acetylated oligomer of different chain length). Thus, the chromatographic system allows the separation of mixtures of partially N-acetylated oligomers according to size (DP) and not according to chemical composition, at least in the separation range between DP 4 (see below for details) and a DP of approximately 20 (Fig. 4; these are the longest oligomers that can be observed separately). It should be noted that the fully acetylated and fully de-N-acetylated standards represent the extreme limits of the possible variation in chemical composition of an oligomer with a given DP, while the enzymatic degradation of chitosans results in oligomer fractions with less variation in chemical composition. As shown in Fig. 1 and by the results presented below, the fully acetylated dimer (AA) and trimer (AAA) are resolved from the corresponding deacetylated dimer (DD) and trimer (DDD). ChiB degradation of a high molecular mass chitosan with F A ¼ 0.65 Time course of the reaction Progress in chitosan degradation can be determined as the degree of scission (a), i.e. the fraction of glycosidic linkages in the chitosan that has been cleaved by the enzyme, by monitoring the increase in reducing end resonances relative to resonances from internal anomer protons in a 1 H-NMR spectrum of the reaction mix- ture. Figure 2 shows the anomer region of the 1 H-NMR-spectrum of a high molecular mass chitosan with F A of 0.65 which had been incubated with ChiB for periods from 15 min (a ¼ 0.03) to 168 h (a ¼ 0.37). These spectra were assigned from previously published assignments of 1 H-NMR spectra of hydro- lysed chitosans [33,38]. The b-anomer reducing end resonance appears as a doublet due to the relatively large coupling constant of 8.1 Hz to H-2. Depending on whether the preceding unit is an A-unit or a D-unit, the b-anomer reducing end resonance appears as two doublets at 4.705 p.p.m. (–AA) and 4.742 p.p.m. (– DA). This assignment was made from 1 H-NMR data on acid hydrolysis of two chitosans with F A of 0.65 and 0.30. The relative intensities of the a ⁄ b-anomers correspond to the equilibrium ratio between the two anomers, which is 60 : 40 at pH 4–5 [39]. The a- and b-anomer reducing end resonances from a D-unit, which would be expected to appear at 5.43 p.p.m. [– D(a)] and 4.92 p.p.m. [–D(b)] [33], are completely absent in the spectrum. The internal D- and A-units resonate at around 4.9 p.p.m. (–D-) and 4.6 p.p.m. (–A-) [9]. Using these assignments, the ratio between internal protons and reducing end protons can easily be calculated. The time course of hydrolysis of the F A ¼ 0.65 chitosan (Fig. 3) shows that the product formation rate is constant until the degree of scission reaches approximately 0.19 (i.e. 19% of the glycosidic linkages has been cleaved). After this initial phase, a second, much slower kinetic phase becomes apparent in which Fig. 1. Calibration of SEC-columns. Degree of polymerization (DP) plotted as a function of the distribution coefficient (K av ) for fully acetylated and fully de-N-acetylated oligomers (monomer to hexamer). The lines correspond to the following equations: log DP ¼ )1.826 K av +1.597 (fully acetylated); log DP ¼ )1.714 K av +1.604 (fully de-N-acetylated). Degradation of chitosans with ChiB A. Sørbotten et al. 540 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS the rate of the enzymatic reaction is decreased by a factor of 4.1. This slow phase continues until the degree of scission reaches 0.37, that is, a situation where on average more than one out of three glycosi- dic linkages have been cleaved (note that for an enzyme producing dimers only, the maximum value of a is 0.5 and that the presence of D-units lowers the maximum value of a). Size distribution of oligomers as a function of a (F A ¼ 0.65) The size distributions of oligomers resulting from hydrolysis of F A ¼ 0.65 chitosan at different a-values are shown in Fig. 4. Undegraded chitosan, with a number average relative molecular mass of 160 000, elutes in the void volume of the column, as do all chitosan chains with DP > 40. Dimer fractions were found to elute in two peaks, which were identified (by 1 H-NMR spectroscopy) as being AA (the dimer with the lowest K av value) and DA. Similarly, trimer and tetramer fractions were found to elute in two peaks. During the initial phase of the degradation process, i.e. at a-values lower than 0.2 (Fig. 3), the peak eluting in the void volume disappeared slowly, while short, mainly even-numbered oligomers (DP 2–12) were pro- duced (Fig. 4). Around a ¼ 0.20, a transition took place: as the void peak disappeared, the reaction slowed down (Fig. 3), and the relative amounts of olig- omers with an odd number of sugar units began to increase. Thus, it seems that the slow phase of the reac- tion starts as the polymer eluting in the void volume is used up. In the final product mixture (a ¼ 0.37), oligo- mer frequencies appear to be determined by oligomer length only, and do not seem to depend on the pres- ence of an odd- or even number of sugar residues. The increased appearance of odd numbered oligomers coin- cides with the appearance of the second, slower kinetic phase visible in Fig. 3. Chemical composition of oligomers To obtain further insight into the chemical composi- tion of the oligomers, we isolated several fractions Fig. 2. Anomer region of the 1 H-NMR spectrum of chitosan (F A ¼ 0.65) after incubation with ChiB for periods ranging from 15 min (a ¼ 0.03) to 10 080 min (a ¼ 0.37) at 37 °C. The reducing end of the a-anomer of an acetylated unit resonates at 5.19 p.p.m., while the b-anomer from the same unit resonates at 4.7–4.8 p.p.m. [33]. D-andA-units within the chain and at nonreducing ends resonate near 4.9 and 4.6 p.p.m., respectively [9]. -AA b and -DA b are the redu- cing end b-anomer where the preceding unit is an A-unit or a D-unit, respectively. Fig. 3. Time course of ChiB degradation of a chitosan with F A ¼ 0.65. The graph shows the degree of scission (a) as a function of time. The slope of the line at low a is 1.8 · 10 )3 min )1 ,andat higher a-values 4.3 · 10 )4 min )1 .Thea -values continue to increase to a ¼ 0.37. A. Sørbotten et al. Degradation of chitosans with ChiB FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 541 obtained at a ¼ 0.11 and a ¼ 0.37, and analyzed these by 1 H-NMR spectroscopy. The yield of odd-num- bered oligomers (a ¼ 0.11) was too low to obtain well-resolved NMR spectra. The results of these analy- ses are summarized in Table 1. Because of an absolute preference for A-units at newly formed reducing ends (i.e. at subsite )1 of the enzyme), only two dimers are formed, AA and DA. In accordance with the NMR analyses shown in Fig. 2, the chromatographic data (Fig. 4) show that, at low a, the dimer mainly consists of AA (86% at a ¼ 0.11). By the end of the reaction, the relative amount of DA starts to increase, finally reaching 34% of the total amount of dimers. NMR analyses of the tetramer fraction at a ¼ 0.11 (Fig. 5A) revealed a F A of 0.73 and a DP of 4.1. Elec- trospray ionization mass spectroscopy (ESI-MS) ana- lysis showed that the dominating tetramer was composed of three A-units and 1 D-unit (results not shown). The minor resonances between 4.8 and 4.9 p.p.m. suggest that minor amounts of DDAA are present. The absence of a doublet at 4.742 p.p.m. in Fig. 5A (–DA; Fig. 2) shows that all tetramers has an A-unit next to the reducing end (–AA). The presence of only one major D-unit resonance at 4.865 p.p.m., and the F A of 0.73 show that there is one dominating tetramer which can be ADAA or DAAA. 1 H-NMR analyses of the tetramer fraction at a ¼ 0.37 revealed a DP of 4.2 and an F A of 0.51 (Fig. 5B). ESI-MS data showed that the majority of the tetramers were composed of two A-units, whereas minor amounts of tetramers with one and three A-units could be detec- ted (results not shown). There are three possibilities for tetramers containing two A-units with an acetylated reducing end, which are DDAA, DADA and ADDA. The fact that at least three different doublet resonances from internal D-units are observed (at 4.855, 4.895 and 4.905 p.p.m.; Fig. 5B) indicates that indeed several oligomers occur. The relative intensities of the doublets at 4.705 p.p.m. (–AA) and 4.742 p.p.m. (–DA) (Fig. 5B) show that the nearest-neighbour to the redu- cing ends are 70% acetylated and 30% deacetylated. Thus, one dominating tetramer must be DDAA. The other possible tetramers with an acetylated reducing end are DDDA, ADDA, DADA, DAAA, ADAA and AADA. Fig. 4. Size-distribution of oligomers after ChiB degradation (F A ¼ 0.65). Chromatograms showing the size-distribution of oligomers obtained after hydrolysis of chitosan (F A ¼ 0.65) with ChiB to degrees of scission (a) varying from 0.03 to 0.37. Peaks are labelled with DP-values or with the sequence of the oligomer. Degradation of chitosans with ChiB A. Sørbotten et al. 542 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS NMR analysis of the trimer fraction (a ¼ 0.37; Fig. 5C) showed a DP of 2.90 and an F A of 0.65. The resonance at 4.705 p.p.m. and the absence of a resonance at 4.742 p.p.m. show that the nearest neighbour to the reducing end unit is an A-unit (–AA). This is confirmed by the two doublet reso- nances at 4.648 and 4.660 p.p.m. with a coupling con- stant of 8.1 Hz, which are assigned to an A-unit next to the reducing end and where the ratio between the two doublets reflects the a ⁄ b-anomer equilibrium at the neighbouring reducing end unit. The resonance at 4.880 p.p.m. is assigned to a deacetylated unit which must be located at the nonreducing end. Thus, the trimer is identified as DAA. Two very small doublet resonances at 4.742 p.p.m. reflect the presence of minor amounts of the trimers with a D-unit next to the acetylated reducing end (i.e. DDA and ⁄ or ADA). Based on the F A -value of 0.65 found for the trimer fraction, relative amounts of the trimers with two acetylated units (DAA and ADA) and the trimer with one acetylated unit (DDA) were calculated to be 95 and 5%, respectively (Table 1). Table 1 also shows F A values for longer oligomers, as determined by 1 H-NMR spectroscopy. The results show that the relative amount of D-units in the longer oligomers increases at high a. This indicates that some D-containing sequences are only hydrolyzed as other, more favourable substrates become scarce. Degradation of chitosans with varying F A Three high molecular mass chitosans with F A ¼ 0.13, 0.32 and 0.50 (Table 2) were incubated with ChiB, and the chitosans were extensively depolymerized to maxi- mum a. The results (Fig. 6) show that the size distribu- tion of the product mixtures shifts towards higher oligomer lengths for substrates with lower F A values and, thus, longer and noncleavable stretches of con- secutive D-units. In addition to a shift to longer oligomers, Fig. 6 shows a reduction in the AA ⁄ DA ratio, as would be expected when lowering the F A of the chitosan. A highly deacetylated chitosan Table 1. Composition of isolated oligomers. Chemical composition and sequence of isolated oligomer fractions obtained after degradation of chitosan (FA ¼ 0.65) with ChiB to degrees of scission (a) of 0.11 and 0.37. ND, not determined. Dimer Trimer Tetramer Pentamer Hexamer Octamer a ¼ 0.11 F A ¼ 0.93 ND F A ¼ 0.73 ND F A ¼ 0.66 F A ¼ 0.61 14% DA ADAA ⁄ DAAA 86% AA a ¼ 0.37 F A ¼ 0.83 F A ¼ 0.65 F A ¼ 0.51 F A ¼ 0.57 F A ¼ 0.53 ND 34% DA DAA DDAA 66% AA DDA ⁄ ADA (ADDA ⁄ DADA ⁄ DAAA ⁄ ADAA ⁄ AADA ⁄ DDDA) Fig. 5. 1 H-NMR spectra (anomer region) of isolated oligomers obtained after hydrolysis of a chitosan with F A ¼ 0.65 with ChiB. (A) Tetramer fraction (a ¼ 0.11). (B) Tetramer fraction (a ¼ 0.37). (C) Trimer fraction (a ¼ 0.37). A. Sørbotten et al. Degradation of chitosans with ChiB FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 543 (F A < 0.001) was not degraded when incubated with an excess of ChiB (result not shown). The chemical compositions of selected oligo- mers (trimer to hexamer) from the experiments with chitosan (F A ¼ 0.32) were investigated by 1 H-NMR spectroscopy, using the methods described above (results not shown). The trimer fraction had a DP of 3.04 and a F A of 0.44, and was concluded to consist mainly of DDA and DAA, in addition to minor amounts of ADA. The tetramer fraction showed a DP of 4.1 and a F A of 0.33 and the dominant tetramers were found to be DDDA and DDAA. 1 H-NMR analyses of the longer oligomers (pentamers, hexamers and heptamers) revealed two dominating oligomers which both contained A-units at the reducing end but which varied with respect to the neighbouring unit (A or D). All other units were deacetylated. Discussion The NMR analysis of reaction products obtained upon degradation of chitosan by ChiB (Figs 2 and 5) showed that all oligomers had acetylated units at their reducing ends. This is in full agreement with the pro- Table 2. Characterization of chitosans. Fraction of acetylated units (F A ), diad frequencies (F AA , F AD , F DA and F DD ), number-average block lengths (N D and N A ) as determined from 600 MHz 1 H-NMR spectroscopy [9]. The intrinsic viscosities ([g]) were determined as described previously [31] and the number-average molecular masses (M n ) were calculated from the intrinsic viscosities [32]. The experimentally deter- mined diad frequencies and block lengths of each chitosan are compared with the calculated diad frequencies and block lengths of a chito- san with a random (Bernoullian) distribution of A- and D-units (numbers given in brackets). ND, not determined. F A [g](mlÆg )1 ) M n F AA F AD ¼ F DA F DD N D N A Chitosan 65% 0.65 740 160 000 0.42(0.42) 0.23(0.23) 0.12(0.12) 1.5(1.5) 2.8(2.9) Chitosan 50% 0.50 850 210 000 0.27(0.25) 0.24(0.25) 0.26(0.25) 2.1(2.0) 2.1(2.0) Chitosan 32% 0.32 820 250 000 0.12(0.10) 0.20(0.22) 0.48(0.46) 3.4(3.1) 1.6(1.5) Chitosan 13% 0.13 910 230 000 ND ND ND ND ND Fig. 6. Size-distribution of oligomers after extended hydrolysis of various chitosans with ChiB. Chromatograms showing the size-distribution of oligomers obtained upon extended ChiB-hydrolysis of chitosans with F A of 0.65, 0.50, 0.32 and 0.13 to a -values (corresponding DP n -val- ues in brackets) of 0.37 (2.7), 0.34 (2.9), 0.22 (4.5) and 0.11 (9.5), respectively. Degradation of chitosans with ChiB A. Sørbotten et al. 544 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS posed substrate-assisted catalytic mechanism of ChiB [15,16]. Whereas many details are known about the catalytic mechanism of ChiB and other family 18 chitinases, much less is known about processivity. Some work has been carried out on ChiA, another family 18 chitinases from S. marcescens which, like ChiB, has a deep active site cleft. It was shown that ChiA is an exoenzyme and degrades the polymer from its reducing end [14,21,28]. On the basis of electron microscopy studies, Uchiyama et al. [28] proposed that ChiA acts processively. In the present study, we have used a new approach to study the properties of ChiB, including possible processivity. When ChiB acts on chitin, it only produ- ces dimers and trimers, indicating that ChiB is an exoenzyme, which can bind either two (subsites )1 and )2) or three (subsites )3, )2 and )1) sugar units in the product binding site [14,21]. The degradation of chitosans yielded products with more than three sugar units very early during the reaction (Fig. 4). This shows that, in the case of a chitosan substrate, the putative physical barrier at the )3 subsite in ChiB [20] does not prevent the enzyme from productive binding events that position more than three sugars on the gly- con side of the catalytic centre. The observation of longer oligomers early during the reaction may seem to contrast with the putative exo-activity of the enzyme, but can be explained by processivity, as dis- cussed below. The slow disappearance of the void vol- ume peak and the early appearance of only shorter oligomers (Fig. 4) indicate that ChiB does indeed degrade the chitosan chains from their ends (as opposed to an endo-action), and thus is an exoenzyme. If ChiB were to be a processive enzyme, one would expect chain movements by an even number of sugar units, as successive glycoside units in chitin ⁄ chitosan are rotated by 180° along the chain, meaning that the cata- lytically important N-acetyl group is positioned cor- rectly in every second sugar only, as proposed for processive cellulases [40–42]. The clear dominance of even-numbered oligomers during the initial phase of the degradation of chitosan with F A 0.65 (Fig. 4) can only be explained by a processive mode of action. If each binding and, for productive binding, cleavage event would be followed by separation and rebinding, the lon- ger initial products would be equally divided between odd-numbered and even-numbered oligomers. The enzyme does not have more than three subsites ()3, )2 and )1) on the glycon side of the catalytic centre, mean- ing that preferential production of, e.g. hexamers and octamers (as opposed to pentamers and heptamers) can- not be explained by specific binding interactions with the enzyme in a nonprocessive mode of action. Taking together all information and observations, ChiB is likely to initially bind the chitosan chain with similar chances of having an odd or even number of sugar units positioned on the glycon side (most likely three, occupying subsites )3to)1, or two, occupying subsites )2 and )1). Depending on the sequence of the bound chain, the first cleavage may then occur directly or after one or more consecutive movements by two sugar units at the time. This would lead to the initial production of an odd- or even-numbered oligomer, respectively. If the enzyme would act processively, all subsequent products coming out of the same binding event will be even-numbered regardless of the initial binding mode, leading to the observed initial domin- ance of even-numbered oligomers. Interestingly, the initial formation of even-numbered oligomers all the way to at least decamers not only reveals the proces- sive mechanism, but also shows that processivity is not interrupted when a nonproductive or less-preferred enzyme–substrate complex emerges (for example a complex that positions a D-unit in the )1 subsite). In the later stages of enzymatic degradation odd-num- bered oligomers appear in larger amounts. This is a consequence of the increased depolymerization of the substrate, which leads to more ‘initial attacks’. It is generally accepted that cellobiohydrolases, which convert cellulose (which is closely related to and equally insoluble as chitin) mainly to dimers, act pro- cessively [40–42]. It is not straightforward to give evi- dence of a processive mechanism, as the parameter that is often used [the dimer ⁄ (trimer + monomer) ratio] strictly speaking does not discriminate between processivity and different initial binding events that occur with particular frequencies, leading to particular dimer ⁄ (trimer + monomer) ratios. The availability of the water-soluble, heteropolymeric chitosan, acting as a ‘pseudo-substrate’ for family 18 chitinases and resembling both chitin and cellulose, allowed us to obtain clear experimental data which unequivocally show that the enzyme acts processively. The current data also provide some insight into the preference for A-orD-units in subsites )2 and +1. Because the nonreducing end sugar of an oligomer must have been bound productively in the +1 subsite in a previous cleavage, the occurrence of both D and A nonreducing ends shows that there is no clear pref- erence for A or D in the +1 subsite. In the initial, fas- ter stage of the depolymerization reaction (a < 0.2), most oligomers formed had an A-unit as nearest neigh- bour to the reducing end (–AA), indicating that most of the productive enzyme–substrate complexes had an A-unit bound in subsite )2. Upon more extensive hydrolysis, presumably leading to depletion of binding A. Sørbotten et al. Degradation of chitosans with ChiB FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 545 sites with the optimal AA sequence in )2 and )1, the reaction becomes slower, and an increase in oligomers with a D-unit as nearest neighbour to the reducing end (–DA) is observed. Thus, ChiB has a clear preference for A in subsite )2. Catalysis in family 18 chitinases requires distortion of the sugar in the )1 subsite [16]. In analogy with, e.g. lysozyme, a certain minimum amount of binding energy is likely to be required to achieve this distortion and formation of a productive enzyme–substrate com- plex. This is reflected by some of the observations made in this study. For example, the trimer AAA is degraded, whereas the trimer DAA (which in principle could bind productively to the )2, )1 and +1 sub- sites) is not. Likewise, several of the tetramers with an –AA reducing end are not degraded while these two A-units could bind productively to the )1 and +1 sub- sites. These observations suggest that the ability to cleave a substrate that positions a D in the )2 subsite depends on how many of the subsites on the aglycone side are occupied by sugars. The distribution of oligomers formed upon extensive ChiB depolymerization of chitosans with varying F A shifted towards higher oligomers with decreasing F A . This was expected because the average chain length of D-blocks that will not be attacked by the chitinase increases with decreasing F A (Table 2). By extensive ChiB-degradation of chitosans with F A from 0.1 to 0.5, we were able to produce longer oligomers (DP > 4) which were composed of only D-units, except for the reducing end unit and in some cases its nearest neighbour. By selecting the chitosan with the appropriate F A and the extent of degradation (a-value), oligomers with predetermined composition and chain length can be prepared. In conclusion, we have found that chitosans with full water-solubility and known random distribution of A- and D-units are useful substrates for obtaining dee- per insight into the mode of action of ChiB, including its processive character. The catalytic mechanism of ChiB requires an acetylated unit in the )1 subsite, but this does not prevent the enzyme from degrading a wide range of chitosans with varying chemical compo- sitions. Thus, ChiB may be used to convert these chitosans to mixtures of oligosaccharides with predict- able chain lengths and chemical compositions. Experimental procedures Chitosans Chitin was isolated from shrimp shells by the method of Hackman [34] and milled in a hammer mill to pass through a 1.0-mm sieve. Chitosans, with degrees of N-acetylation of 65, 50, 32 and 13% (F A ¼ 0.65, 0.50, 0.32 and 0.13), were prepared by homogeneous de-N-acetylation of chitin [35]. The chemical properties of the chitosans (F A , diad frequen- cies, DP n ) were characterized by 1 H-NMR spectroscopy [5,9], and by intrinsic viscosity measurements [36]. Number- average molecular masses were calculated from the Mark– Houwink–Sakurada equation, as reported by Anthonsen et al. [37]. Characteristic features of the chitosans are given in Table 2, including the fraction of acetylated units (F A ), the diad frequencies, the intrinsic viscosities ([g]), the num- ber-average molecular mass (M n ) and the average length of the D-blocks (N D ). These characteristics show that the chitosans have a random distribution of acetylated units, i.e. according to Bernoullian distribution. Chitinase B ChiB was overexpressed in Escherichia coli and purified by a protocol consisting of a previously described hydrophobic interaction chromatography step [20,21], preceded by ion- exchange chromatography on Q-Sepharose Fast Flow (Amersham Pharmacia Biotech AB, Uppsala, Sweden). The final protein material was collected in an ammonium carbon- ate buffer [21], and freeze-dried. Before use, the enzyme was dissolved to 1 mgÆmL )1 in 20 mm trizma base, pH 7.5. Pro- tein concentrations were determined with the Bio-Rad Pro- tein Assay (Bio-Rad Laboratories Inc., Hercules, CA, USA). The purified protein used in this work displayed the same specific activity and kinetic parameters as ChiB used in previ- ous studies, including the protein used for successful X-ray diffraction studies [20]. Kinetics of chitosan degradation Samples of 10 mg chitosan (F A ¼ 0.65) were dissolved in 1.0 mL H 2 O and added to 1.0 mL 0.08 m NaAc buffer, pH 5.5, containing 0.2 m NaCl and 0.2 mg BSA. The sam- ples were placed in a shaking waterbath at 37 °C. The depolymerization reaction was started by adding 5 lgof ChiB and was stopped after between 15 and 10 080 min by adjusting the pH to 2.5 with 1.0 m HCl, and immersing the samples in boiling water for 2 min. The samples were stored at )18 °C until further analysis. Extended depolymerization of chitosans with varying degree of acetylation (F A ) Four samples of 10 mg chitosan (F A ¼ 0.13, 0.32, 0.50 and 0.65) were each dissolved in 1.0 mL H 2 O and added to 1.0 mL 0.08 m NaAc buffer, pH 5.5, containing 0.2 m NaCl and 0.2 mg BSA. The samples were immersed in a shaking waterbath at 37 °C. The depolymerization reactions were started by adding 10 lg of ChiB and the reactions were stopped after 1 week as described above. Degradation of chitosans with ChiB A. Sørbotten et al. 546 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 1 H-NMR-spectroscopy The samples were dissolved in D 2 O and the pD was adjus- ted to 4 with DCl. The deuterium resonance was used as a field-frequency lock, and the chemical shifts were refer- enced to internal sodium 3-(trimethylsilyl)propionate-d 4 (0.00 p.p.m.). The 1 H-NMR spectra were obtained at 90 °C at 300.13, 400.13 or 600.13 MHz as previously described [5,9], and F A values were calculated as described by Va ˚ rum et al. [9]. The number average degree of polymerization, DP n , of enzyme-degraded chitosans was determined from the anomer (H-1) resonances as follows: DP n ¼ [area of all H-1-resonances (internal and reducing end)] ⁄ (area of redu- cing end resonances). The degree of scission, a, was calcula- ted as a ¼ 1 ⁄ DP n . Size-exclusion chromatography (SEC) Oligomers from enzymatic depolymerization reactions were separated on three columns in series, packed with Superdexä 30, from Amersham Pharmacia Biotech (overall dimensions 2.60 · 180 cm). The column was eluted with 0.15 m ammonium acetate, pH 4.5 at a flow rate of 0.8 mLÆmin )1 . The effluent was monitored with an online refractive index (RI) detector (Shimadzu RID 6 A, Shimadzu Schweiz GmbH, Reinach, Switzerland), coupled to a datalogger. Fractions of 3.2 mL were collected for analysis. Standard samples contained 10 mg of partially depolymerized chitosan. In experiments where oligomers were collected for further analysis by NMR spectroscopy, samples containing up to 200 mg of partially depolymerized chitosan were injected without loss of resolution. Fully acetylated and fully de-N-acetylated oligomer were used as standards (monomer to hexamer; Seikagaku Corporation, Tokyo, Japan). To verify the possibility to quantify amounts of oligosac- charides with varying DP and F A from the refractive index detector response, we checked the detector response in the following way: exact amounts (1 mg) of the fully acetylated and fully de-N-acetylated oligomers (dimer, tetramer and hexamer) were dissolved in the mobile phase and injected on the columns, and the refractive index detector signal was used to determine the area of the peaks after elution from the column. These analyses showed that, within the accuracy of the area determination, there was a linear rela- tionship between peak areas and the amount (mass) of injected oligomer, irrespective of DP and degree of acetyla- tion. Acknowledgements This work was supported by grants from the Norwegian Research Council (140497 ⁄ 420 and 134674 ⁄ I10). We thank Hege Grindheim for carrying out some of the experiments. References 1 Roberts GAF (1992) Chitin Chemistry, 1st edn. Macmil- lan, London. 2Va ˚ rum KM, Ottøy MH & Smidsrød O (1994) Water- solubility of partially N-acetylated chitosans as a function of pH – effect of chemical composition and depolymerization. Carbohydr Polym 25, 65–70. 3 Kristiansen A, Va ˚ rum KM & Grasdalen H (1998) The interactions between highly de-N-acetylated chitosans and lysozyme from chicken egg white studied by 1 H-NMR spectroscopy. Eur J Biochem 251, 335–342. 4 Nordtveit RJ, Va ˚ rum KM & Smidsrød O (1994) Degra- dation of fully water-soluble, partially N-acetylated chito- sans with lysozyme. Carbohydr Polym 23, 253–260. 5Va ˚ rum KM, Holme HK, Izume M, Stokke BT & Smidsrød O (1996) Determination of enzymatic hydro- lysis specificity in partially N-acetylated chitosans. Biochim Biophys Acta 1291, 5–15. 6 Schipper NGM, Va ˚ rum KM & Artursson P (1996) Chitosans as absorption enhancers for poorly absorb- able drugs.1. Influence of molecular weight and degree of acetylation on drug transport across human intestinal epithelial (Caco-2) cells. Pharmaceut Res 13, 1686–1692. 7Ko ¨ ping-Ho ¨ gga ˚ rd M, Tubulekas I, Guan H, Edwards K, Nilsson M, Va ˚ rum KM & Artursson P (2001) Chitosan as a nonviral gene delivery system: structure–property relationships and characteristics compared with poly- ethylenimine in vitro and after lung administration in vivo. Gene Ther 8, 1108–1121. 8 Kurita K, Sannan T & Iwakura Y (1977) Evidence for formation of block and random copolymers of N-acetyl- d-glucosamine and d-glucosamine by heterogeneous and homogeneous hydrolyses. Makromol Chem 178, 3197– 3202. 9Va ˚ rum KM, Anthonsen MW, Grasdalen H & Smidsrød O (1991) Determination of the degree of N-acetylation and the distribution of N-acetyl groups in partially N-deacetylated chitins (chitosans) by high-field NMR- spectroscopy. Carbohydr Res 211, 17–23. 10 Va ˚ rum KM, Anthonsen MW, Grasdalen H & Smidsrød O (1991) 13 C-NMR studies of the acetylation sequences in partially N-deacetylated chitins (chitosans). Carbo- hydr Res 217, 19–27. 11 Coutinho PM & Henrissat B (1999) Carbohydrate- Active Enzymes server. http://afmb.cnrs-mrs.fr/$cazy/ CAZY/index.html 12 Watanabe T, Kimura K, Sumiya T, Nikaidou N, Suzuki K, Suzuki M, Taiyoji M, Ferrer S & Regue M (1997) Genetic analysis of the chitinase system of Serratia mar- cescens 2170. J Bacteriol 79 , 7111–7117. 13 Suzuki K, Taiyoji M, Sugawara N, Nikaidou N, Henris- sat B & Watanabe T (1999) The third chitinase gene (ChiC) of Serratia marcescens 2170 and the relationship A. Sørbotten et al. Degradation of chitosans with ChiB FEBS Journal 272 (2005) 538–549 ª 2005 FEBS 547 [...].. .Degradation < /b> of < /b> chitosans < /b> with < /b> ChiB 14 15 16 17 18 19 20 21 22 23 24 25 26 of < /b> its product to other bacterial chitinases Biochem J 343, 587–596 Suzuki K, Sugawara N, Suzuki M, Uchiyama T, Katouno F, Nikaidou N & Watanabe T (2002) Chitinases A, B and C1 of < /b> Serratia marcescens 2170 produced by recombinant Escherichia coli: enzymatic properties and synergism on chitin degradation < /b> Biosci Biotechnol Biochem... Comparative studies of < /b> chitinases A and B from Serratia marcescens Microbiology 142, 1581–1589 Robyt JF & French D (1967) Multiple attack hypothesis of < /b> alpha-amylase action–action of < /b> porcine pancreatic human salivary and Aspergillus oryzae alpha-amylases Arch Biochem Biophys 122, 8–16 Robyt JF & French D (1970) Multiple attack and polarity of < /b> action of < /b> porcine pacreatic alpha-amylase Arch Biochem Biophys 138,... (2004) Mutational and computational analysis of < /b> the role of < /b> conserved residues in the active site of < /b> a family 18 chitinase < /b> Eur J Biochem 271, 253–262 van Aalten DMF, Synstad B, Brurberg MB, Hough E, Riise BW, Eijsink VGH & Wierenga RK (2000) Structure of < /b> a two-domain chitotriosidase from Serratia marcescens at 1.9-angstrom resolution Proc Natl Acad Sci USA 97, 5842–5847 Brurberg MB, Nes IF & Eijsink... I, Stahlberg J, Koivula A, Piens K, Becker D, Boer H, Harle R, Harris M, Divne C, Mahdi S, Zhao Y, Driguez H, Claeyssens M, Sinnott ML & Teeri TT (2003) Engineering the exo-loop of < /b> Trichoderma reesei cellobiohydrolase, Cel7A: a comparison with < /b> Phanerochaete chrysosporium Cel7D J Mol Biol 333, 817–829 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS Degradation < /b> of < /b> chitosans < /b> with < /b> ChiB 42 Varrot A, Frandsen... crystalline chitin hydrolysis by Chitinase < /b> A from Serratia marcescens 2170 J Biol Chem 276, 41343–41349 ˚ 29 Vander P, Varum KM, Domard A, El Gueddari NE & Moerschbacher BM (1998) Comparison of < /b> the ability of < /b> partially N-acetylated chitosans < /b> and chitooligosaccharides to elicit resistance reactions in wheat leaves Plant Physiol 118, 1353–1359 30 Akiyama K, Kawazu K & Kobayashi A (1995) Partially N-deacetylated... (1992) Chitosan cross-linked with < /b> Mo(VI) polyoxyanions – a new gelling system Biomaterials 13, 635–638 ˚ 37 Anthonsen MW, Varum KM & Smidsrød O (1993) Solution properties of < /b> chitosans:< /b> conformation and chain stiffness of < /b> chitosans < /b> with < /b> different degrees of < /b> N-acetylation Carbohydr Polym 22, 193–201 ˚ 38 Varum KM, Ottøy MH & Smidsrød O (2001) Acid hydrolysis of < /b> chitosans < /b> Carbohydr Polym 46, 89–98 39 Tsukada... Wilson KS & Dijkstra BW (1997) Substrate-assisted catalysis unifies two families of < /b> chitinolytic enzymes J Am Chem Soc 119, 7954–7959 ˚ van Aalten DMF, Komander D, Synstad B, Gaseidnes S, Peter MG & Eijsink VGH (2001) Structural insights into the catalytic mechanism of < /b> a family 18 exo -chitinase < /b> Proc Natl Acad Sci USA 98, 8979–8984 Mark BL, Vocadlo DJ, Knapp S, Triggs-Raine BL, Withers SG & James MNG... evidence for substrate-assisted catalysis in a bacterial betahexosaminidase J Biol Chem 276, 10330–10337 Williams SJ, Mark BL, Vocadlo DJ, James MNG & Withers SG (2002) Aspartate 313 in the Streptomyces plicatus hexosaminidase plays a critical role in substrateassisted catalysis by orienting the 2-acetamido group and stabilizing the transition state J Biol Chem 277, 40055–40065 Synstad B, Gaseidnes S,... Davies G & Henrissat B (1995) Structures and mechanisms of < /b> glycosyl hydrolases Structures 3, 853–859 Rouvinen J, Bergfors T, Teeri T, Knowles JKC & Jones TA (1990) Three-dimensional structure of < /b> cellobiohydrolase-II from Trichoderma reesei Science 249, 380–386 Divne C, Stahlberg J & Reinikainen T (1994) The 3dimensional crystal-structure of < /b> the catalytic core of < /b> cellobiohydrolase-I from Trichoderma reesei... Conformational properties of < /b> chito-oligosaccharides: titration, optical rotation and carbon-13 NMR studies of < /b> chito-oligosaccharides Carbohydr Res 88, 19–38 40 Teeri TT, Koivula A, Linder M, Wohlfahrt G, Divne C & Jones TA (1998) Trichoderma reesei cellobiohydrolases: why so efficient on crystalline cellulose? Biochem Soc Trans 26, 173–178 FEBS Journal 272 (2005) 538–549 ª 2005 FEBS A Sørbotten et al 41 von . Degradation of chitosans with chitinase B from Serratia marcescens Production of chito-oligosaccharides and insight into enzyme processivity Audun Sørbotten 1 ,. were started by adding 10 lg of ChiB and the reactions were stopped after 1 week as described above. Degradation of chitosans with ChiB A. Sørbotten et al. 546 FEBS