An ATIII–Sepharose column (30 × 7 cm) prepared by coupling 2 g of human ATIII to CNBr-activated Sepharose 4B (Sigma) as described by Höök and coworkers (32) was used. A step gradient of NaCl concentration was applied. Low-affinity fractions were eluted using a 0.25 M NaCl solution buffered at pH 7.4 with 1 mM Tris–HCl at 11 mL/min; high-affinity fractions were eluted by a five-step gradient of NaCl (0.71, 1.39, 2.07, 2.72, and 3.5 M NaCl and 1 mM Tris–HCl, pH 7.4). The NaCl gradient was monitored by conductivity measurements, and heparin fractions were detected by ultraviolet (UV) light at 219 nm. This wavelength limits the influence of any NaCl present in the mobile phase and is sensitive to N-acetyl functional groups across the heparin chain. Chromatograms were obtained after subtraction of the signal obtained using a blank run. Injected quantities varied between 250 and 500 mg of heparin, depending on the capacity of the column and the type of heparin.

Multistep desalting was necessary for the elimination of NaCl, especially for fractions of highest affinity. First, high-affinity fractions (1L) were diluted 1:5 in water and passed through a 20 × 1.6 cm column filled with Q-Sepharose Fast Flow (Sigma Aldrich, Saint-Quentin-Fallavier, France). The column was then washed with water to eliminate free NaCl. The heparin was then flash-eluted by NaClO₄ 2.5 N. UV detection at 215 nm was used to monitor the elution. In a second step, the heparin solution was desalted on a 100 × 7 cm column filled with Sephadex G10 monitored with UV detection at 215 nm and conductimetry.

Standard Carbopack AS11 conditions used a single 250 × 2.1 mm column (Thermo Scientific Dionex, Montigny-le-Bretonneux, France). The column temperature was set at 40°C. Mobile phase A was 2.5 mM NaH₂PO₄ at pH 2.8, and mobile phase B was an aqueous solution of 2.5 mM NaH₂PO₄ with 1 M NaClO₄ adjusted to pH 3.0. A linear gradient (t_0min B% 0; t_80min B% 60) was applied for elution at a flow rate of 0.22 mL/min. Diode array detection was used. Double UV detection was performed at 232 nm and 202–247 nm.

The liquid chromatography/mass spectrometry (LC/MS) method used to analyze heparin digests was as described previously (8), and used an Acquity UPLC BEH C18 column, 150 × 2.1 mm, 1.7 μm (Waters SAS, En Yvelines Cedex, France). Mobile phase A was water, and mobile phase B water:acetonitrile (30:70). The ion pairing reagent, heptyl amine (HPTA; 7.5 mM), and a buffering agent, hexafluroisopropanol (50 mM), were added to both A and B. A linear gradient (t_0min B% 1; t_70min B% 70) was applied for elution at a flow rate of 0.22 mL/min. Column temperature was set at 30°C and diode array detection used. Double UV detection was performed at 265 nm and 232 nm. Electrospray ionization mass spectra were obtained using a Waters Xevo Q-Tof mass spectrometer. The electrospray interface was set in negative ion mode with a capillary potential of 2000 V and a sampling cone potential of 20 V. Source and desolvation temperatures were 120 and 300°C, respectively. Nitrogen was used as the desolvation (750 L/min) and cone (25 L/min) gas. The mass range was 50–2500 Da (scan rate = 0.8 s). Acquisition was performed in MSE mode with low energy at 7 V and a high energy ramp from 30 to 50 V.

The sulfanilic tagged heparin digests (100 μL) were diluted 1/3 to 1/5 in 5 mM Na₂HPO₄ at pH 7.0 before being treated for 24 h at room temperature with 20 milliunits of Δ⁴⁻⁵ 2-O-sulfatase or Δ⁴⁻⁵ glycuronidase.

To demonstrate that unsaturated 3S disaccharides are generated by heparinase I digestion, a BMH was digested with a heparinases I+II+III mixture and with heparinase II alone (Figure 1). In a preliminary statement, especially important when the selectivity of heparinase is concerned, it should be emphasized that heparinases from Flavobacterium heparinum were used in this work. We previously observed that heparinases obtained as recombinant enzymes may differ in activity level and specificity from their naturally generated counterparts, which could in turn influence the results of any equivalent analysis. In a similar way, heparinases from Bacteroides eggerthii (33) (New England Biolabs, Ipswich, MA) have demonstrated different substrate specificities compared with those from Flavobacterium heparinum.

Heparinase I preferentially cleaves highly sulfated zones and has a primary selectivity for GlcNS (± 6S)1 → 4IdoA(2S) linkages (20–22) with a pure endolitic mechanism (Supplementary Figure 1). By contrast, heparinase II has broader selectivity (20–22) but its mechanism of action is more complex. This enzyme has two separate actions with a high probability of the presence of two active sites (34), one acting endolitically and the other exolitically (Supplementary Figure 2). The processive exolytic mechanism is acting at the reducing end (unpublished results). However, the major characteristic of heparinase II in relation to this work is that it cannot cleave the non-reducing link of the 3S disaccharidic units (12, 14–17). In a more general way, despite its broad selectivity, the cleavage of low sulfated disaccharides such as ΔIVs (ΔHexUA-GlcNS) or even ΔIVa (ΔHexUA-GlcNAc) appears difficult, if not impossible, in sequences specifically cleaved by heparinase III. Heparinase III cleaves low-sulfated sequences, and therefore has no selective action on 3S moieties. The main heparinase III specificity is directed at the glycoserine bond ^↓GlcA β1-3 Gal β1-3 Gal β1-4 Xyl β1- (31), and its absence from the lyase mixture is demonstrated by the presence in the digest of complete glycoserine moieties, with no effect on 3S building blocks. BMH is the only heparin in which ΔIIIs is present in significant amounts (8), and three unsaturated 3S disaccharides were accordingly detected in the chromatogram after heparinase I+II+III digestion. However, as shown in Figure 1 (chromatogram b), the depolymerization of 3S units into disaccharides is not possible using heparinase II since the NRS is no longer cleavable, with the digestion fragments eluting as higher oligosaccharides, where the 3S disaccharides are found at the RE, as in ΔIs-Is_id [(ΔHexUA(2S)-GlcN(NS,6S)-IdoA(2S)-GlcN(NS,3S,6S); see Table 2, lines 7 and 8] (Full structural identification in Supplementary Table 1).

Chromatograms of the same heparin digested by heparinase I are shown in Figure 2. The digest is more complex than those shown in Figure 1, and gel permeation chromatography (GPC) was used to facilitate identification of components. In contrast to heparinase II, 3-O-sulfation facilitates rather than hinders cleavage on the non-reducing linkage. The three 3S disaccharides in Figure 2 are present at the same levels as found using mixed heparinases (Figure 1A), although only heparinase I was used. This is even the case for ΔIIs where in the absence of uronic 2-O-sulfate, the cleavage of the non-reducing linkage is considerably more difficult.

ΔIIs-IIs_glu [ΔHexUA-GlcN(NS,6S)-GlcA-GlcN(NS,3S,6S)] and ΔIs-IIs_glu [ΔHexUA(2S)-GlcN(NS,6S)-GlcA-GlcN(NS,3S,6S)] are also present (Figure 2). ΔIIs-IIs_glu appears as a result of cleavage of -Is↓idIIsglu-IIs_glu, which predominates over -Is↓idIIsid-IIs_glu (note that the two hexasaccharides ΔIs-IIs_glu-IIs_glu [(ΔHexUA(2S)-GlcN(NS,6S)-GlcA-GlcN(NS,6S))-GlcA-GlcN(NS,3S,6S)] and ΔIs-IIs_id-IIs_glu [(ΔHexUA(2S)-GlcN(NS,6S)-IdoA-GlcN(NS,6S))-GlcA-GlcN(NS,3S,6S)] have been isolated from PMH after heparinase I digestion in our laboratory [unpublished results]). Given that ΔIs-IIs_glu [ΔHexUA(2S)-GlcN(NS,6S)-GlcA-GlcN(NS,6S)] (14) appears in the digest at reaction completion as a consequence of resistance to cleavage of the -Is↓idIIsglu linkage (Table 2, lines 1 and 2), the generation by heparinase I of ΔIIs-IIs_glu is difficult to account for. The reason why -Is↓idIIsglu-IIs_glu is cleaved when -Is↓idIIsglu is resistant can be explained by the novel observation of the distance of the link from the RE. In modeled reactions on key decasaccharides [ΔIs-Is_id-Is_id-Is_id-Is_id, ΔIs-Is_id-Is_id-Is_id-IIIs_id, ΔIs-Is_id-Is_id-Is_id-IIs_glu (unpublished results)], progressive resistance to cleavage is observed as the bond moves closer to the RE.

The fact that resistance to cleavage can be increased approximately 4-fold when the link is directly on the RE is of major importance, especially when sequencing oligosaccharides purified from LMWH (35). Ernst et al. (19) highlighted this effect on key decasaccharides, but concluded incorrectly that the mechanism was exolytic. There is now an overall consensus across the literature (14, 36, 37) that the behavior of heparinase I is endolytic, as evident from GPC of heparin digests (14, 38). The influence of distance from the RE is essential for understanding the depolymerization mechanism of heparinase I, and for explaining why hexasaccharides like ΔIs-IIa_id-IIs_glu, are resistant, not because of, but despite, their 3-O-sulfation. In the heparinase I digests, prominent selectivity for cleavage of -GlcN(NS,3S,6S)^↓IdoA(2S) linkages (14) is due to the highly sulfated environment. Under these conditions, endolytic cleavage first targets ATIII-binding pentasaccharides and generates 3S reducing glucosamines. For the main PMH binding site, -Is_id-IIa_id-IIs_glu [-IdoA(2S)-GlcN(NS,6S)-IdoA-GlcNAc(6S)-GlcA-GlcN(NS,3S,6S); Table 2, line 5], the ^↓ IIs_glu (^↓GlcA-GlcN(NS,3S,6S)) link cannot be cleaved because of (i) its position on the RE and (ii) low sulfation directly on both sides of the non-reducing link. Furthermore, the absence of 2-O-sulfation, and importantly the N-acetylation of the glucosamine moiety, inhibits -Is↓idIIsid-IIs_glu cleavage so that the hexasaccharide ΔIs-IIa_id-IIs_glu is obtained (15). As indicated in Figure 2, when the glucosamine is N-sulfated, cleavage is possible, and a tetrasaccharide is obtained (case of ΔIIs-IIs_glu and ΔIs-IIs_glu). For N-acetylated pentasaccharides, the cleavage of the NRS link is possible only where 2-O-sulfation is present, as in -Is_id-Ia_id-IIs_glu (-IdoA(2S)-GlcN(NS,6S)-IdoA(2S)-GlcNAc(6S)-GlcA-GlcN(NS,3S,6S), yielding the tetrasaccharide ΔIa-IIs_glu (Table 2, lines 1 and 4). In all these cases, 3-O-sulfation triggers only the first cleavage, which places the 3S glucosamine at the RE. Note that cleavage selectivity of heparinase I for this new sequence, containing the 3S glucosamine on the RE, is no longer dependent on the reducing 3-O-sulfate, and identical cleavage would occur on a similar sequence without any 3S.

In the digestion by heparinase I of key decasaccharides, cleavage of intermediate tetrasaccharides (ΔIs-Is_id for ΔIs-Is_id-Is_id-Is_id-Is_id) (ΔIs-Is_id and ΔIs-IIIs_id for ΔIs-Is_id-Is_id-Is_id-IIIs_id) starts only after the elimination of all other longer fragments as a result of decreasing lyase activity at the RE. In general, tetrasaccharide digestion by heparinase I is possible only when sulfation in positions 2, 6 and N is present. If the glucosamine is 6-OH, as in ΔIs-IIIs_id, digestion is very difficult. It is also blocked when the uronic acid is 2-OH, especially with glucuronic configurations as in ΔIs-IIs_glu. Therefore, the tetrasaccharide residues of ATIII-binding sites, even the most heavily sulfated (ΔIs-IIs_glu), cannot be reduced to disaccharides (or to only a very small extent). The only exception is when the unsaturated disaccharide is 3S as in ΔIIs-IIs_glu [ΔHexUA-GlcN(NS,3S,6S)-GlcA-GlcN(NS,3S,6S)], which is at least partially digested into ΔIIs depending on the quantity of heparinase I added (Supplementary Figure 6).

The presence in heparinase I digests of the 3S disaccharides ΔIs, ΔIIs and ΔIIIs (Figure 2) raises the question of the structural parameters which allow complete digestion to these products. In the case of ΔIs and ΔIIIs, 2-O-sulfation of the uronic acid is the key factor: when associated with 3-O-sulfation of the glucosamine residue, NRS cleavage is observed whatever the configuration of the uronic acid. The question of the presence of ΔIIs in heparin digests is less straightforward. If we consider results obtained with synthetic 3S hexasaccharides (18), unsaturated 3S disaccharides would always be obtained when the glucosamine on the NRS of the 3S unit is N- and 6-O-sulfated, since the cleavage by heparinase I is possible in all proposed examples of that type. However, these observations cannot be directly applied to 3S sequences in heparin chains because the position of the 3S glucosamines in this work are not on the RE, as in digested heparin. Instead, the 3S glucosamines are in the medium position of the hexasaccharides, on the NRS of the reducing disaccharide, which induces specific heparinase I behavior.

Similarly, in sequencing experiments on oligosaccharides purified from high-affinity fractions of LMWH [enoxaparin (39) and semuloparin (40)] carried out at our laboratory, the formation of ΔIIs from -IIs_glu- was frequently observed when the glucosamine on the NRS was N-sulfated. These experiments started with a short oligosaccharide (hexasaccharide to decasaccharide) containing the entire ATIII binding site. Consequently, the cleavage on -IIs_glu^↓ by heparinase I was frequently hindered by simple inhibition at the reducing link or by another reducing moiety resistant to heparinase I such as a mannosamine or a 1,6-anhydro ring (39). Under these circumstances, heparinase I first cleaved the non-reducing link^↓ IIs_glu-, resulting ultimately in the formation of ΔIIs.

This behavior is due to specific structures generated by the LMWH depolymerization process, but these structural features could not be found (or were present only in trace amounts) in a heparin partially digested by heparinases. In such cases, the 3S glucosamines are predominantly situated at REs after preliminary cleavage (as described previously), and heparinase-resistant 3S tetrasaccharides are obtained. When the specificity of heparinase I is considered, it is difficult to rule out possible switching to an iduronic configuration, i.e., -IIs_id [-IdoA-GlcN(NS,3S,6S)], which could probably allow ΔIIs transformation, or at least facilitate it. This has been reported previously by Mochizuki et al. (17). IIs_id could be obtained from IIs_id with the presence of 3-O-sulfotransferases (3-OST) (24, 41): ΔIIs has been detected (23, 26) in 3-OST-modified heparan sulfate. Thus, the possibility to obtain ΔIIs from IIs_id must be examined.

However, several factors militate against the IIs_id hypothesis: first, ΔIIs is absent in ATIII low-affinity heparin fractions, and increases with affinity of the heparin fraction for ATIII (see later). Therefore, ΔIIs cannot be classified as a non-ATIII binding 3-O-sulfate, as in Mochizuki et al. (17). Furthermore, it appears that IIs_id is not compatible with the structural requirements of ATIII binding sequences and sequences containing IIs_id have never been found to bind ATIII (18). Additionally, in nitrous depolymerizations of heparins, the transformed building block of IIs_id, IdoA-aMan3S6S (in which aMan represents 2,5-anhydromannose), has never been detected (42). Indeed, we believe that ΔIIs results from digestion by heparinase I of sequences with consecutive 3-O-sulfated units.

In the next section, where digestion by heparinase mixture and by heparinase II alone are compared, ATIII binding sites with a supplementary 3-O-sulfate such as -IIa_id-IIs_glu-Is_id- [IdoA-GlcNAc(6S)-GlcA-GlcN(NS,3S,6S)-IdoA(2S)-GlcN(NS,3S,6S)] are found in heparin fractions with highest affinity for ATIII. They are detected in heparinase II digests as unsaturated hexasaccharides ΔIIa-IIs_glu-Is_id. ΔIIa-IIs_glu-Is_id-Is_id was identified (43) in our laboratory in structural studies on the high ATIII affinity octasaccharide fraction from semuloparin, which contains this unusual 3S element. Furthermore, the same sequence had already been reported in ATIII high-affinity fractions of BMH (27). Based on this accumulating evidence, we believe that heparin contains other types of consecutive 3S units, including -IIs_glu-. Several examples have been isolated in ATIII high-affinity fractions of PMH in our laboratory. The full structural identification of one of these, the decasaccharide ΔIs-IIa_id-IIs_glu-IIs_glu-IIs_glu [ΔHexUA(2S)-GlcN(NS,6S)-IdoA-GlcNAc(6S)-GlcA-GlcN(NS,3S,6S)-GlcA-GlcN(NS,3S,6S)-GlcA-GlcN(NS,3S,6S), is shown in Supplementary Table 2, followed by its sequencing via fractionation with heparinase in Supplementary Figures 6, 7]. In the presence of heparinase II, a mixture of ΔIIa-IIs_glu-IIs_glu-IIs_glu and ΔIs is obtained. Digestion by heparinase I first gives a mixture of ΔIs-IIa_id-IIs_glu, ΔIIs-IIs_glu and ΔIIs, and then ΔIs-IIa_id-IIs_glu and ΔIIs only at completion.

ATIII affinity chromatography was used to fractionate heparins into an ATIII low-affinity fraction (LA) and five fractions of increasing affinity for ATIII (HA1 to HA5) (Figure 4). A five-step concentration gradient (0.71, 1.39, 2.07, 2.72, 3.5 M NaCl) with injection of ~500 mg of heparin was used. The properties of BMH and collected fractions are summarized in Table 3.

Unidentified building blocks were detected in fractions from BMH (same batch as analyzed in Figure 3). The main tetrasaccharides in Figure 3B-2 (other than the five basic 3S tetrasaccharides ΔIIa-IIs_glu, ΔIIa-IVs_glu, ΔIIs-IIs_glu, ΔIa-IIs_glu, ΔIs-IIs_glu) correspond to one ΔU(4,3,0) (peak 1), two ΔU(4,4,0) (peaks 2 and 3), five ΔU(4,5,0) (peaks 4 to 6, ΔIIIs-IIs_glu (ΔHexUA(2S)-GlcNS-GlcA-GlcN(NS,3S,6S), ΔIs-IVs_glu [ΔHexUA(2S)-GlcN(NS,6S)-GlcA-GlcN(NS,3S)] and three ΔU(4,6,0) (peaks 7 to 9). The LC/MS ion pair chromatogram digests (heparinase I+II+III and II alone) of LA, HA1, HA3 and HA5 from BMH with RIC corresponding to ΔU(4,3,0), ΔU(4,4,0), ΔU(4,5,0) and ΔU(4,6,0) are shown in Figures 5–8. Peaks ΔU(4,5,0)-4, ΔU(4,6,0)-7, ΔU(4,6,0)-8 and ΔU(4,6,0)-9 are not specific to BMH, and were also present in BLH digests. The remainder appear to be linked to 6-O-desulfation, so that their content (estimated with UV at 265 nm using LC/MS ion pair chromatography) is much lower in heparin sources other than BMH (Table 4). Δ⁴⁻⁵ glycuronidase and Δ⁴⁻⁵ 2-O sulfatase were applied to all digests to limit possible search options, and their action on each tetrasaccharide is also included in Table 4.

The constituents of exhaustive digests from BMH ATIII affinity fractions have already been described by Naggi et al. (27, 44). Four supplementary 3S tetrasaccharide building blocks (two with molecular weight Mw 994 Da [ΔU(4,4,0)] and two at 1074 Da [ΔU(4,5,0)] were detected in this work. Structures tentatively proposed were ΔIIs-IVs_glu [ΔHexUA-GlcN(NS,6S)-GlcA-GlcN(NS,3S)] and ΔIIIs-IVs_glu [ΔHexUA(2S)-GlcNS-GlcA-GlcN(NS,3S)] for the two at 994 Da, and ΔIIIs-IIs_glu and ΔIs-IVs_glu for those at 1074 Da. In accordance with Naggi et al. (27, 44), ΔIIIs-IIs_glu and ΔIs-IVs_glu appear to be appropriate choices for two of the five ΔU(4,5,0) unknowns in light of their distribution in the affinity fractions and their transformation by Δ⁴⁻⁵ glycuronidase and Δ⁴⁻⁵ 2-O-sulfatase.

Among the tetrasaccharides listed in Table 4, five were sensitive to Δ⁴⁻⁵ 2-O-sulfatase: the two tetrasaccharides identified as ΔIIIs-IIs_glu and ΔIs-IVs_glu, ΔIs-IIs_glu, and ΔU(4,5,0)-5 and ΔU(4,6,0)-7. The 6-O-sulfate on the non-reducing glucosamine of the ATIII pentasaccharide binding site is a major contributor to the affinity for ATIII (45). It corresponds to the one in the unsaturated unit of the tetrasaccharides, so that the presence in LA fractions of ΔIIIs-IIs_glu as a major ΔU(4,5,0) component is easily explained by the 6-O-desulfation of ΔIIIs-. Low amounts (<0.1%) of the ΔU(4,3,0) tetrasaccharide ΔIVs-IVs_glu [ΔHexUA-GlcNS-GlcA-GlcN(NS,3S)] (initially present in trace amounts) were also obtained when Δ⁴⁻⁵ 2-O-sulfatase was added. This was due to digestion of ΔIIIs-IVs_glu as proposed in (44) as a possible ΔU(4,4,0). However, additionally, peaks ΔU(4,4,0)-2 and ΔU(4,4,0)-3 increased as a result of the enzymatic Δ⁴⁻⁵ 2-O-desulfation of ΔIIIs-IIs_glu and ΔIs-IVs_glu. It thus seems logical to propose for these two peaks the structures ΔIVs-IIs_glu [ΔHexUA-GlcNS-GlcA-GlcN(NS,3S,6S)] and ΔIIs-IVs_glu with a coelution (probably for peak ΔU(4,4,0)-2) reflecting the additional presence of ΔIIIs-IVs_glu.

Among the species shown in Table 4, only ΔIIs-IIs_glu was fully digested by Δ⁴⁻⁵ glycuronidase, while a partial sensitivity was observed for tetrasaccharide ΔU(4,4,0)-2, with the concomitant formation of a trisaccharide, G(3,4,0) (Mw 836Da). As Δ⁴⁻⁵ glycuronidase has stronger reactivity on 6-O-sulfated substrates (46), peak ΔU(4,4,0)-2 could be due to ΔIIs-IVs_glu, and then, consequently, peak ΔU(4,4,0)-3 due to ΔIVs-IIs_glu. In contrast to hexasaccharide building blocks, complete or partial inhibition of Δ⁴⁻⁵ glycuronidase was frequently observed in the tetrasaccharides. Absence of reaction to Δ⁴⁻⁵ 2-O-sulfatase thus appears a better indicator of a Δ2-OH uronic acid than sensitivity to Δ⁴⁻⁵ glycuronidase. Another simple way to check Δ2-O-sulfatation of a component can be found in the chromatographs of the digests before sulfanilic tagging as shown in Figures 1, 2 through its UV spectrum by the 2 nm hypsochromic shift in the 232 nm maximum of UV absorbance in the case of a 2-O-sulfated unsaturated acid (6). The presence of a Δ2-OH uronic acid was thus ensured for tetrasaccharides ΔU(4,5,0)-4, ΔU(4,6,0)-8 and ΔU(4,6,0)-9, in line with the absence of reaction to Δ⁴⁻⁵ 2-O-sulfatase. The action of Δ⁴⁻⁵ 2-O-sulfatase on a digest of BLH fraction HA1 was used to confirm that tetrasaccharide ΔU(4,5,0)-4 is Δ2-O-desulfated tetrasaccharide ΔU(4,6,0)-7, as per the tetrasaccharides ΔU(4,4,0)-2 and ΔU(4,4,0)-3 corresponding to the Δ2-O-desulfated tetrasaccharides ΔIs-IVs_glu and ΔIIIs-IIs_glu, respectively.

Tetrasaccharides ΔU(4,6,0)-8 and ΔU(4,6,0)-9 are however a special case, in that they are detected preferentially in HA4 and HA5 fractions. In addition, when digestion is limited to heparinase II, they disappear and are shifted to hexasaccharides as ΔIIs and ΔIIIs are to tetrasaccharides. Thus, the corresponding unsaturated disaccharide should also be 3-O-sulfated, and the heparinase II-corresponding hexasaccharide building blocks ΔU(6,8,0), ΔU(6,9,0) or ΔU(6,7,1) in the case of N-acetylation. ΔIIs-IIs_glu appears to be an ideal candidate, even if the full digestion of ΔIIs-IIs_glu by heparinase I was observed in the sequencing experiment of ΔIs-IIa_id-IIs_glu-IIs_glu-IIs_glu (Supplementary Figure 6). The good correlation observed between the percentages of ΔU(4,6,0)-8 and ΔU(4,6,0)-9 with the percentage of ΔIIs, in BMH and BLH, argues in favor of their partial depolymerization into ΔIIs with some ΔIIs-IIs_glu remaining and corresponding to ΔU(4,6,0)-8. It should be noted that two ΔU(4,6,0) tetrasaccharides other than ΔIs-IIs_glu have already been reported in BLH (47).