MALDI-TOF MS and ESI-LTQ-Orbitrap tandem mass spectrometry reveal specific porphyranase activity from a Pseudoalteromonas atlantica bacterial extract

Abstract

A better understanding of the chemical–physical properties of porphyran – a complex anionic polysaccharide – and the development of potential industrial applications requires more in-depth knowledge of its structural organization. The structural complexity of the hybrid (co-polymer) structure of porphyran stems from the co-occurrence of two repetition moieties LA-G and L6S-G, which can be methylated on the D-galactose. The enzymes currently available for the structural analyses of porphyran are limited in specificity, and methylated porphyran cannot be digested by previously described β-porphyranases. Here, MALDI-TOF MS analysis and tandem ESI-MS sequencing of porphyran degradation products after incubation with protein extracts from the marine bacterium Pseudoalteromonas atlantica revealed methylated disaccharide (L6S-GMe) and dimethylated tetrasaccharide (L6S-GMe–L6S-GMe) end-products that have never been described before. Our results highlighted unprecedented β-methyl-porphyranase activity that can accommodate the methylated building blocks of porphyran.

---

1. Introduction

Porphyran is a water-soluble agaran belonging to the sulfated galactans family, which also includes carrageenans.^1–4 Sulfated galactans are linear sulfated polysaccharides extracted from the cell wall of red algae and are widely used for industrial applications in the food, cosmetic, and pharmaceutical sectors.^1,5–13 Like carrageenans, agarans are hybrid polysaccharides composed of several disaccharide repetition moieties, but they differ stereoisometrically: the galactan backbone of carrageenans ((D-G)_n) is made exclusively of D-galactose residues alternately linked by α(1 → 3) and β(1 → 4) glycosidic bonds, whereas in agarans ((L-G)_n), the α-linked galactose belongs to the L-series.^14–16 The disaccharide repetition moieties can vary in the substitution occurring on the free hydroxyl groups. These substitutions are generally ester sulfate (S) and a 3,6 anhydro-ring (DA or LA) on the 4-linked residue.^14,17,18 Other substitutions such methyl (Me) or pyruvate groups (P) also increase the number of possible carrageenan or agaran structures.^17–19 For example, agarose refers to a polysaccharide composed essentially of LA-G (agarobiose) repetition moieties, but also contain methylated and sulfated moieties, whose amounts depend on the algal source and the method of extraction.²⁰

Porphyran is extracted from the cell wall of red algae belonging to the genus Porphyra, which includes one of the most highly consumed red algal species.²¹ The structural complexity of this polysaccharide is due to the co-occurrence of two repetition moieties LA-G and L6S-G, which can be methylated on the [β-linked] D-galactose giving rise two other repetition moieties: LA-GMe and L6S-GMe (Scheme 1).^17,22 LA-G is formed in vivo from the biosynthetic precursor L6S-G via galactose-6-sulfurylase, and it can also be produced in vitro by alkaline treatment of L6S-G.^23,24 The L6S-G moieties are very abundant in porphyran and are also present in most untreated agarans.

Scheme 1 Chemical structure of the disaccharide building blocks occurring in porphyran. G = D-galactose, G6Me = 6-O-methyl-D-galactose, L = L-galactose, L6S = 6-O-sulfate-L-galactose, LA = 3,6-anhydro-L-galactose.

The hybrid (co-polymer) structure of porphyran was first demonstrated by enzymatic degradation with β-agarase^25–27 and more recently with β-porphyranase A.¹⁷ Agarases specifically target the LA-G moieties, which represent about 30% of porphyran^25,28 whereas β-porphyranases accommodate sulfated L6S-G moieties in their active site.^17,29 Degradation of agarose repetition units is well documented and numerous enzymes leading to complete saccharification have been described.^30–32 These enzymes include α- and β-agarases, α-3,6-anhydrogalactosidase and β-galactosidase. In contrast, only a few enzymes that digest segments of L6S-G units have been observed and biochemically characterized. β-Porphyranase A and B from the marine bacterium Zobellia galactinivorans belonging to the GH16 family of glycoside hydrolases (http://www.cazy.org) have been studied by enzymology and crystallography thereby depicting the sub-site topology of the active site as well as their substrate specificity.²⁹ In particular, methylated porphyran cannot be digested by β-porphyranases suggesting that enzymatic digestion of the methylated fraction of porphyran requires other enzymes. Compared to the diversity of other classes of carbohydrate-active enzymes such as the κ-, ι- and λ-carrageenases, the enzymes available for the structural analysis of porphyran are limited in specificity.

Recent genomic and metagenomic investigations of a large variety of biological sources indicate that there is a huge untapped reservoir of new polysaccharide-degrading enzymes, especially porphyranases.³³ Investigation of gene clusters in Z. galactanivorans, which correspond to polysaccharide-utilizing loci likely dedicated to the degradation of marine polysaccharides, have led to the identification of new porphyranases.³⁴ Interestingly, proteins homologous to these bacterial porphyranases have been found in Japanese samples of the human intestinal microbiome. These enzymes were probably acquired after gene transfer from the marine strain present in food made with Porphyra sp. (e.g. sushi) through horizontal gene transfer from marine bacteria to human gut bacteria.³³

We previously implemented a screening method adapted to the detection of polysaccharide-degrading enzymes in complex protein extracts from various marine bacterial strains.³⁵ Combined with mass spectrometry (MS), this screening method is a powerful tool for the structural analysis of carbohydrate degradation products.³⁶ Here, we report detailed MS analyses of degradation products obtained after incubation of porphyran with protein extracts from the marine bacterium Pseudoalteromonas atlantica (strain T6c).³⁷ The detailed MS characterization of the resulting oligosaccharides revealed unprecedented series of oligo-porphyran end-products that resulted from heretofore undocumented porphyranase activity.

2. Experimental

2.1 Materials

2-(4-Hydroxyphenylazo)benzoic acid (HABA) and 1,1,3,3-tetramethylguanidine (TMG) were purchased from Sigma-Aldrich Co. (Saint-Quentin-Fallavier, France). HPLC-Grade methanol was obtained from VWR (West Chester, PA, USA). Water was of ultrapure quality, obtained from a MilliQ apparatus (Millipore, Milford, USA). Porphyran was extracted from the red alga Porphyra umbilicalis and showed the same chemical features as those previously published.¹⁷

2.2 Porphyran depolymerization by purified β-porphyranase A

Porphyran oligosaccharides were obtained by the enzymatic degradation of P. umbilicalis porphyran catalyzed by the recombinant β-porphyranase A from Z. galactanivorans expressed in Escherichia coli,³⁸ according to a previously described protocol.¹⁷ Briefly, purified porphyran (1% w/v in deionized water) was incubated for 12 h at 30 °C with pure β-porphyranase A (0.15 μg mL⁻¹). The completion of the enzyme reaction was ensured by the addition of extra β-porphyranase A (0.15 μg mL⁻¹). The reaction was stopped by heat inactivation (10 min) of the enzyme. Progress of enzymatic depolymerization was monitored by means of reducing-end measurements using a method adapted from Kidby and Davidson,³⁹ and analytical SEC as previously described.¹⁷ Porphyran oligosaccharides were purified by preparative SEC using three Superdex 30 (26/60) (GE Healthcare) columns in series, integrated in a high-performance liquid chromatography (HPLC) system liquid injector/collector and monitored by Unipoint software (Gilson). The freeze-dried hydrolysis product was dissolved in deionized water at 4% (w/v), and filtered through a 0.45 μm syringe filter (Millipore). Then, 4 mL samples were loaded onto Superdex 30 columns. Detection was carried out by refractive index measurement (Spectra System RI-50) and 50 mM ammonium carbonate [(NH₄)₂CO₃] was used as a running buffer at 1.5 mL min⁻¹ flow rate. Fractions of oligosaccharides were pooled according to their degree of polymerization and subsequently freeze-dried. Oligosaccharide samples were diluted ten-fold either in water for matrix-assisted laser desorption/ionization-time of flight MS (MALDI-TOF MS) analysis or in 1 : 1 water : methanol (v/v) for electrospray ionization MS (ESI-MS) analysis.

2.3 Porphyran depolymerization by a protein extract from Pseudoalteromonas atlantica

Pseudoalteromonas atlantica T6c was grown at 18 °C in Zobell medium.⁴⁰ A preculture was carried out by inoculating frozen cells in 10 mL of Zobell medium in a 100 mL Erlenmeyer flask. After incubation for 36 h at 18 °C and then shaking at 150 rpm in a New Brunswick incubator, 1 mL of the culture was inoculated in 50 mL of fresh Zobell medium and incubated for a further 8 h under the same conditions. Then, 50 mL was inoculated in 1 L of Zobell medium under the same conditions (18 °C, 150 rpm, 5 L shaker flask). After 48 h, the culture was centrifuged at 5000 × g for 30 min at 4 °C. Culture medium was collected and underwent several filtration steps as described in Fer et al.³⁵ First, filtration through a 300 kDa cut-off membrane was carried out to remove high-molecular-weight macromolecules and cell debris. Then, the filtrate was diafiltered on a 10 kDa cut-off membrane using 100 mM Tris buffer pH 7.5 to eliminate low-molecular-weight molecules (e.g. free sugars). The protein extracts, which contained molecules of molecular weight comprised between 10 kDa and 300 kDa, were incubated with porphyran (0.1% w/v in deionized water) for 20 h at 34 °C. The reaction mixture was then filtered through a 10 kDa cut-off membrane to separate and recover low-molecular-weight oligosaccharides in the filtrate. Oligosaccharide samples were diluted two-fold either in water for MALDI-TOF MS analysis or in 1 : 1 water : methanol (v/v) for ESI-MS analysis.

2.4 MS analysis

MALDI-TOF MS experiments were performed using a PerSeptive Biosystems Voyager-DE Pro STR MALDI-TOF mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA). This instrument was equipped with a nitrogen UV laser (λ = 337 nm) pulsed at a 20 Hz frequency. The mass spectrometer was operated in both the positive and negative ion reflector mode with an accelerating potential of 20 kV and a grid percentage equal to 70%. Mass spectra were recorded with the laser intensity set just above the ionization threshold (2700 in arbitrary units, on our instrument) to avoid fragmentation and labile group losses, to maximize the resolution (pulse width 3 ns) and to obtain a strong analyte signal with minimal matrix interference. Time delay between laser pulse and ion extraction was set to 300 ns. Mass spectra were typically obtained by accumulation of 200–300 laser shots in reflector mode and processed using Data Explorer 4.0 software (Applied Biosystems). The HABA/TMG₂ ionic liquid used as the matrix was prepared as described elsewhere.^41–43 Briefly, HABA was mixed with TMG at a 1 : 2 molar ratio in methanol. The solution was then sonicated for 15 min at 40 °C. After removal of methanol by centrifugal evaporation in a SpeedVac for 3 h at room temperature, the ionic liquid matrix (ILM) was left under vacuum overnight. Final solutions were then prepared at a concentration of 90 mg mL⁻¹ in methanol and used as a matrix without further purification. Once prepared, these solutions of ILMs can be stored at 4 °C up to 1 week. A 1 : 1 (v/v) mixture of oligosaccharide sample : ILM was prepared, and 1 μL of the mixture was loaded on a polished stainless steel MALDI target and allowed to dry at room temperature and atmospheric pressure for 5 min.

ESI-MS experiments were carried out on a LTQ-Orbitrap XL from Thermo Scientific (San Jose, CA, USA) and operated in negative ionization mode, with a spray voltage at −5 kV. A 1 : 1 water : methanol (v/v) mixture was continuously infused using a 250 μL syringe at 10 μL.min⁻¹ flow. Applied voltages were −35 and −110 V for the ion transfer capillary and the tube lens, respectively. The ion transfer capillary was held at 275 °C. Resolution was set to 60 000 (at m/z 400) for all experiments, and the m/z ranges were set to 300–2000 in profile mode and in the normal mass range during full-scan experiments. Spectra were analyzed using the acquisition software XCalibur 2.0.7 (Thermo Scientific, San Jose, CA, USA), without smoothing or background subtracts. During MS/MS scans, collision-induced dissociation (CID) fragmentation occurred in the linear ion-trap analyzer and detection in Orbitrap with centroid mode. For CID fragmentation, an activation Q value (Q) of 0.25 and an activation time (T) equal to 30 ms were used. Normalized collision energy (NCE) was set at 25% for MS/MS experiments and 15–25% for MSⁿ experiments (with n > 2). The automatic gain control (AGC) allowed accumulation up to 1 × 10⁶ ions for Fourier-transform MS (FTMS) scans, 2 × 10⁵ ions for FTMSⁿ scans and 1 × 10⁴ ions for ion-trap MS (ITMSⁿ) scans. Maximum injection time was set to 500 ms for both FTMS and FTMSⁿ scans and 100 ms for ITMSⁿ scans. For all scan modes, 1 μscan was acquired. The precursor selection window was 2 Da and 3 Da during MS² and MSⁿ (with n > 2), respectively. Each ion with z ≥ 2 present in the full-scan MS and exhibiting a signal intensity of >500 arbitrary units was fragmented by iterative ESI-MSⁿ steps (with n ≥ 2). Fragments resulting from fragmentation steps were annotated according to the nomenclature described by Domon and Costello.⁴⁴

3. Results and discussion

3.1 Comparison of β-porphyranase A and P. atlantica end-products

After incubation of a protein extract from the marine bacterium P. atlantica with porphyran, MALDI-TOF MS analysis revealed a series of oligosaccharides, thereby indicating that P. atlantica possesses porphyranase activity (Fig. 1A). The spectrum of the incubation showed a wide distribution of ions ranging from m/z 421.07 (dp2) to m/z 2815.47 (dp14) (ESI, Table S1†) revealing a new, undocumented series of oligo-porphyrans. Mass spectrum of the heat shock inactivated P. atlantica extract clearly revealed no ions corresponding to oligosaccharides (data not shown). For comparison, the MALDI mass spectrum of the depolymerization products obtained after digestion of porphyran with purified β-porphyranase A (Fig. 1B) exhibited fewer species ranging from m/z 421.07 (dp2) to m/z 2283.49 (dp12) (ESI, Table S2†). The di- and tetrasaccharide end-products formed by β-porphyranase A have been previously characterized by NMR and showed that the disaccharide α-L-Galp-6-sulfate (1 → 3) β-D-Galp (L6S-G) is the major end-product and that higher oligosaccharides belong to methylated and non-methylated L6S-G series.^17,29 Accordingly, the three most intense ions in the mass spectrum of the β-porphyranase A end-products were at m/z 421.07, 877.11 and 1169.12 (Fig. 1B) corresponding to the disaccharide L6S-G, a methylated tetrasaccharide (L6S0-G)₂ + CH₃ and a disulfated hexasaccharide including an anhydrogalactose residue (LA-G)(L6S-G)₂, respectively. The first striking difference between P. atlantica and β-porphyranase A end-products was observed in the range of dp2. The L6S-G (m/z 421.07) was the only dp2 limit product produced by β-porphyranase A, whereas the P. atlantica extract yielded two dp2 limit products at m/z 421.07 (L6S-G) and at m/z 435.09 (Fig. 1A). The latter dp2 ion was the most intense and was ascribed to a monomethylated disaccharide (L6S-G + CH₃). For higher-molecular-weight oligosaccharides, the number of species was systematically greater for P. atlantica extracts than for β-porphyranase A end-products. For example, the dp4 species detected after β-porphyranase A digestion included two oligosaccharides without and with one methyl group (m/z at 863.08 and 877.11, respectively) and was based on a (L6S-G)₂ backbone (Fig. 1B), while at least three monosodiated tetrasaccharides with no (m/z 847.11), one (m/z 861.12) and two (m/z 875.13) methyl groups were observed with the P. atlantica extract (Fig. 1A). Similarly, the dp6 oligosaccharides detected after β-porphyranase A digestion were based on a [(LA-G)(L6S-G)₂ − 2Na + K]⁻ backbone with no, one or two methyl groups, while the more diverse backbones ([(LA-G)(L-G)(L6S-G)]⁻, [(L-G)₂(L6S-G)]⁻, [(LA-G)(L6S-G)₂], [(L-G)(L6S-G)₂]) were observed for the dp6 oligosaccharides formed by the P. atlantica extracts (ESI, Table S1 and S2†).

Fig. 1 Negative ion reflector MALDI-TOF mass spectra of porphyran depolymerization reaction mixtures incubated with (A) Pseudoalteromonas atlantica protein extract and (B) β-porphyranase A. Fully detailed annotations of the different ions are given in ESI.†

3.2 Sequencing of β-porphyranase A oligosaccharides using ESI-MSⁿ

The β-porphyranase A depolymerization products have been previously analyzed by NMR,¹⁷ but they have only been partly characterized by mass spectrometry using a particularly high-energy photon activation,³⁶ and the location of the methyl groups and the discrimination between several isomers had remained unknown. In this study, end-products of β-porphyranase A were purified by SEC and analyzed using MALDI-TOF revealing slight size heterogeneity in SEC fractions (Table 1, and ESI Text S1 and Fig. S1†), and ESI MSⁿ in negative ionization. ESI-MSⁿ can determine characteristic MSⁿ fragmentation patterns that were then used to elucidate the new series of oligo-porphyrans produced by the P. atlantica extract (ESI Text S2 and Fig. S2† for a complete description of the MSⁿ experiments).

Table 1 Ions detected by MALDI-TOF MS in negative mode after incubation of β-porphyranase A with porphyran and separation by SEC (only ions with ≥1% of the relative abundance are reported). All ions detected within each SEC fraction were listed, and putative oligosaccharide composition (not the sequence) is given

Fractions	m/z ([M − Na]^−)		Mass accuracy (ppm)	Putative composition
	Experimental	Theoretical
dp2	421.10	421.07	24	[(L6S-G) − Na]^−
dp4	583.15	583.15	0	[(G)(L6S-G) − Na]^−
	597.13	597.13	0	[(G)(L6S-G) + CH3 − Na]^−
	745.17	745.17	0	[(L-G)(L6S-G) − Na]^−
	759.19	759.19	0	[(L-G)(L6S-G) + CH3 − Na]^−
	847.11	847.11	0	[(L6S-G)2 − Na]^−
	861.14	861.13	12	[(L6S-G)2 + CH3 − Na]^−
	863.08	863.08	0	[(L6S-G)2 − 2Na + K]^−
	877.10	877.10	0	[(L6S-G)2 + CH3 − 2Na + K]^−
	1033.25	1033.26	10	[(LA-G)2(L6S-G) − Na]^−
	1051.32	1051.31	10	[(LA-G)(L-G)(L6S-G) − Na]^−
	1153.22	1153.20	17	[(LA-G)(L6S-G)2 − Na]^−
dp4 − Me	597.13	597.13	0	[(G)(L6S-G) + CH3 − Na]^−
	699.06	699.07	14	[(L6S-G)(L6S) + CH3 − Na]^−
	745.17	745.17	0	[(L-G)(L6S-G) − Na]^−
	759.19	759.19	0	[(L-G)(L6S-G) + CH3 − Na]^−
	847.11	847.11	0	[(L6S-G)2 − Na]^−
	861.13	861.13	0	[(L6S-G)2 + CH3 − Na]^−
	863.08	863.08	0	[(L6S-G)2 − 2Na + K]^−
	877.11	877.10	11	[(L6S-G)2 + CH3 − 2Na + K]^−
	1051.32	1051.31	10	[(LA-G)(L-G)(L6S-G) − Na]^−
	1065.30	1065.28	19	[(LA-G)(L-G)(L6S-G) + CH3 − Na]^−
	1153.21	1153.20	9	[(LA-G)(L6S-G)2 − Na]^−
	1167.23	1167.22	9	[(LA-G)(L6S-G)2 + CH3 − Na]^−
dp6	745.19	745.17	27	[(L-G)(L6S-G) − Na]^−
	759.21	759.19	26	[(L-G)(L6S-G) + CH3 − Na ]^−
	861.15	861.13	23	[(L6S-G)2 + CH3 − Na]^−
	1051.31	1051.31	0	[(LA-G)(L-G)(L6S-G) − Na]^−
	1065.28	1065.28	0	[(LA-G)(L-G)(L6S-G) + CH3 − Na]^−
	1153.20	1153.20	0	[(LA-G)(L6S-G)2 − Na]^−
	1167.22	1167.22	0	[(LA-G)(L6S-G)2 + CH3 − Na]^−
	1183.20	1183.19	8	[(LA-G)(L6S-G)2 + CH3 − 2Na + K]^−
dp6 − Me	1051.30	1051.31	10	[(LA-G)(L-G)(L6S-G)2 − N]^−
	1065.28	1065.28	0	[(LA-G)(L-G)(L6S-G) + CH3 − Na]^−
	1079.29	1079.30	9	[(LA-G)(L-G)(L6S-G) + 2CH3 − Na]^−
	1083.29	1083.29	0	[(L-G)2(L6S-G) + CH3 − Na]^−
	1153.20	1153.20	0	[(LA-G)(L6S-G)2 − Na]^−
	1167.22	1167.22	0	[(LA-G)(L6S-G)2 + CH3 − Na]^−
	1181.25	1181.24	8	[(LA-G)(L6S-G)2 + 2CH3 − Na]^−
	1183.17	1183.19	17	[(LA-G)(L6S-G)2 + CH3 − 2Na + K]^−
	1185.21	1185.23	17	[(L-G)(L6S-G)2 + CH3 − Na]^−
	1201.19	1201.20	8	[(L-G)(L6S-G)2 + CH3 − 2Na + K]^−
	1213.25	1213.26	8	[(L-G)(L6S-G)2 + 3CH3 − Na]^−
	1215.21	1215.22	8	[(L-G)(L6S-G)2 + 2CH3 − 2Na + K]^−
	1303.14	1303.14	0	[(L6S-G)3 + CH3 − 2Na + K]^−
	1317.18	1317.16	15	[(L6S-G)3 + 2CH3 − 2Na + K]^−
	1357.34	1357.36	15	[(L-G)(LA-G)2(L6S-G) − Na]^−
	1371.36	1371.38	15	[(L-G)(LA-G)2(L6S-G) + CH3 − Na]^−
	1375.38	1375.37	7	[(LA-G)(L-G)2(L6S-G) − Na]^−
	1459.32	1459.30	14	[(LA-G)2(L6S-G)2 − Na]^−
	1473.33	1473.32	7	[(LA-G)2(L6S-G)2 + CH3 − Na]^−

---

Fragmentation of the dp2 ion at m/z 421.0645 led to the formation of a non-reducing end sulfated galactose and fragment ions resulting from intra-cyclic cleavages that established the presence of a sulfate group at the C-6 position of the non-reducing end galactose (Fig. S2A†). These MS/MS data confirmed that the dp2 fraction was the monosulfated disaccharide galactosyl-galactose L6S-G, in full agreement with the typical building block depicted for porphyran.¹⁷ This L6S-G disaccharide appeared to be the limit product of the enzymatic β-porphyranase A depolymerization of porphyran.

The tetrasaccharide SEC fraction contained two disulfated tetrasaccharide species detected as doubly charged ions at m/z 412.0583 and m/z 419.0663, the latter having a mass increment of 14.0160 mass units corresponding to a methyl group (data not shown). The ESI-MS² fragmentation of the disulfated tetrasaccharide ion at m/z 412.0583 resulted in cross-ring cleavages and glycosidic bond cleavages leading to non-reducing end mono-, di- and trisaccharides and reducing-end trisaccharides, which together with further MS³ fragmentations (data not shown) were in agreement with the following porphyran tetrasaccharide sequence: (L6S-G)–(L6S-G) (ESI, Fig. S2B†). The methylated disulfated tetrasaccharide at m/z 419.0663 also underwent ESI-MS² and MS³ fragmentation, in both negative and positive ionizations, and located the methyl group on the second galactose residue from the non-reducing end in the following tetrasaccharide sequence: (L6S-G6Me)–(L6S-G) (ESI, Fig. S2C†).

Similar MSⁿ investigations were conducted with the four detected dp6 end-products of β-porphyranase A, which shared a common (L6S-G)–(LA-G)–(L6S-G) backbone but varied according to the occurrence of 0, 1 or 2 methyl groups. Fragmentation of the non-methylated hexasaccharide with its doubly charged ion [M − 2Na]²⁻ at m/z 565.1051 led to the following sequence: (L6S-G)–(LA-G)–(L6S-G) localizing the anhydro-galactose (LA) in the central disaccharide (ESI, Fig. S2D†). Fragmentation of the doubly charged ion [M − 2Na]²⁻ at m/z 572.1131 corresponding to a monomethylated hexasaccharide revealed that this dp6 ion could be attributed to two hexasaccharide isomers differing in the position of the methyl group. Further MSⁿ fragmentations (data not shown) located the methyl group on the G unit either in the internal LA-G disaccharide or in the non-reducing end LA-G disaccharide, leading to the following complete sequences of the two isomers: (L6S-G)–(LA-G6Me)–(L6S-G) and (L6S-G6Me)–(LA-G)–(L6S-G) (ESI, Fig. S2E†). Taking into account the intensity of specific fragment ions for each isomer, we assumed a ≈ 30 : 70 ratio for (L6S-G6Me) : (LA-G6Me). With regard to the dimethylated hexasaccharide detected at m/z 579.1204, its MS² fragments indicated that the methyl groups were not located at the reducing-end disaccharide. Further fragmentations delineated the following sequence: (L6S-G6Me)–(LA-G6Me)–(L6S-G) for this fourth hexasaccharide isomer contained in the dp6 fraction (ESI, Fig. S2F†).

3.3 Characterization of a new series of methylated oligo-porphyrans produced by P. atlantica

The ESI-MS analysis of the depolymerization products from the P. atlantica incubation confirmed the presence of two dp2 limit products as non-methylated and methylated disaccharides detected at m/z 421.0635 and at m/z 435.0792, respectively (Inset, Fig. 2). The ESI-MS² fragmentation of the methylated dp2 ion [M − Na]⁻ at m/z 435.0792 in negative ionization mode, as well as in positive ionization mode for its counterpart [M + H]⁺ positive ion at m/z 459.0801, located the sulfate group and the methyl moiety at the C-6 of the L-galactose (L) unit and at the C-6 of the reducing D-galactose residue (G), respectively (data not shown), leading to the dp2 structure L6S-G6Me.

Fig. 2 (A) Negative ESI-MS spectrum of porphyran depolymerization reaction mixture incubated with Pseudoalteromonas atlantica protein extract. Inset: close-up of the dp2 m/z area pointing to the disaccharide limit products at m/z 421.0645 and 435.0799, without and with a methyl group, respectively.

Higher dp oligosaccharides showed higher structural diversity than those obtained with β-porphyranase A (Table 2). In addition to the oligo-porphyran dp4 and dp6 already described with β-porphyranase A, new distinct oligosaccharides were detected and characterized. In particular, we observed the occurrence of a dimethylated dp4 as a discharged ion at m/z 426.0750 (Fig. 2). MS/MS experiments yielded a major B₃²⁻ ion at m/z 329.0355 corresponding to a disulfated trisaccharide with only one methyl remaining (Fig. 3A). It resulted from the glycosidic bond cleavage that released the reducing-end methylated galactose from the parent tetrasaccharide and formed the B₃²⁻ ion [M_dp4–G6Me − 2Na]²⁻. Two cross-ring fragments at m/z 308.0311 (^0,2A₃³⁻) and m/z 374.0510 (^1,4A₄²⁻) confirmed the presence of a methyl group on the C-6 of the reducing-end galactose. Further MS³ fragmentation experiments targeting the B₃²⁻ ion at m/z 329.0355 provided an intense peak at m/z 417.0674 (Fig. 3B) that corresponded to a monosulfated disaccharide including one methyl group. The peak at m/z 417.0674 resulted from the loss of a L6S moiety and was mainly attributed to a B₂ fragment corresponding to a [M_dp4–(L6S-G6Me) − Na]⁻ species. This m/z value can also contain a small part of Y₃, but B₂ was favored due to preferential charge localization at the non-reducing end. Other glycosidic bond cleavages were also observed, leading to B₁/C₁ ions at 241.0015/259.0023 and C₂ ion at 435.0788. Two minor cross-ring cleavages at m/z 138.9706 (^0,4A₁) and m/z 299.0257 (^2,4X₃²⁻), attested to the presence of a sulfate group on the L-galactose of each disaccharide repeat. Further fragmentations via an MS⁴ step on the most intense B₂ at m/z 417.0674 led to a dominant ion at m/z 241.0022 assigned to a B₁ fragment (Fig. 3C) corresponding to a sulfated galactose. Taken together, these fragments confirmed the methylation of the G residue located in the third position from reducing end of the tetrasaccharide. MS⁵ fragmentation of the B₁ ion (m/z 241.0022) led to the loss of sulfate anion at m/z 96.9603, and to cross-ring cleavages ^0,4A₁ (m/z 138.9708), ^0,3A₁-H₂O (m/z 150.9705), ^0,3A₁ (m/z 168.9809) and ^0,4A₁-H₂O (m/z 180.9808), which unambiguously located the sulfate group at the C-6 of the non-reducing-end galactose. Overall, these iterative MS fragmentations were consistent with the following tetrasaccharide sequence: (L6S-G6Me)–(L6S-G6Me). In addition to a higher degree of methylation, the P. atlantica oligosaccharides also showed differences in their methylation pattern. Unlike β-porphyranase A oligosaccharides, a significant part of the detected tetrasaccharides formed by P. atlantica were found methylated on the reducing-end galactose (Table 2).

Table 2 Sequences of oligosaccharides formed after incubation with Pseudoalteromonas atlantica extract and determined by ESI-MSⁿ in negative mode (only ions with ≥1% of the relative abundance are reported). All oligosaccharides have been submitted to the same procedure

dp	m/z ([M − nNa]^n−)		Mass accuracy (ppm)	Deduced sequences
	Experimental	Theoretical
a Number of isomers ≥3.b Number of isomers ≥4.
2	421.0645^1−	421.0652^1−	1.7	L6S-G
	435.0799^1−	435.0809^1−	2.3	L6S-G6Me
4	727.1594^1−	727.1603^1−	1.2	LA-G–L6S-G
	741.1750^1−	741.1759^1−	1.2	LA-G6Me–L6S-G (18%)
				LA-G–L6S-G6Me (82%)
	755.1903^1−	755.1916^1−	1.7	LA-G6Me–L6S-G6Me
	412.0593^2−	412.0599^2−	1.5	L6S-G–L6S-G
	419.0672^2−	419.0677^2−	1.2	L6S-G6Me–L6S-G (35%)
				L6S-G–L6S-G6Me (65%)
	426.0750^2−	426.0756^2−	1.4	L6S-G6Me–L6S-G6Me
6	1033.2549^1−	1033.2554^1−	0.5	L6S-G–LA-G–LA-G
	1047.2703^1−	1047.2710^1−	0.7	L6S-G6Me–LA-G–LA-G (45%)
				L6S-G–LA-G6Me–LA-G (30%)
				L6S-G–LA-G–LA-G6Me (25%)
	1061.2861^1−	1061.2867^1−	0.6	L6S-G6Me–LA-G6Me–LA-G (56%)
				L6S-G6Me–LA-G–LA-G6Me (31%)
				L6S-G–LA-G6Me–LA-G6Me (13%)
	565.1069^2−	565.1075^2−	1.1	L6S-G–LA-G–LA-G
	572.1146^2−	572.1153^2−	1.2	L6S-G6Me–LA-G–L6S-G (39%)
				L6S-G–LA-G6Me–L6S-G (31%)
				L6S-G–LA-G–L6S-G6Me (30%)
	579.1225^2−	579.1231^2−	1.0	L6S-G6Me–LA-G6Me–L6S-G (71%)
				L6S-G6Me–LA-G–L6S-G6Me (22%)
				L6S-G–LA-G6Me–L6S-G6Me (7%)
	409.0582^3−	409.0582^3−	0.0	L6S-G–L6S-G–L6S-G
	413.7297^3−	413.7300^3−	0.7	L6S-G6Me–L6S-G–L6S-G (57%)
	—	—		L6S-G–L6S-G6Me–L6S-G (30%)
	—	—		L6S-G–L6S-G–L6S-G6Me (13%)
	418.4012^3−	418.4019^3−	1.7	L6S-G6Me–L6S-G6Me–L6S-G (79%)
	—	—		L6S-G6Me–L6S-G–L6S-G6Me (17%)
	—	—		L6S-G–L6S-G6Me–L6S-G6Me (4%)
	423.0732^3−	423.0738^3−	1.4	L6S-G6Me–L6S-G6Me–L6S-G6Me
10	484.0812^4−	484.0811^4−	0.2	(L6S-G)4(LA-G)
	487.5847^4−	487.5849^4−	0.4	(L6S-G)4(LA-G) + 1Me^a
	491.0886^4−	491.0888^4−	0.4	(L6S-G)4(LA-G) + 2Me^a
	494.5923^4−	494.5928^4−	1.0	(L6S-G)4(LA-G) + 3Me^a
	498.0962^4−	498.0967^4−	1.0	(L6S-G)4(LA-G) + 4Me^a
12	567.1172^4−	567.1178^4−	1.1	(L6S-G)4(LA-G)2 + 3Me
	574.1255^4−	574.1257^4−	0.4	(L6S-G)4(LA-G)2 + 5Me
	470.2817^5−	470.2830^5−	2.8	(L6S-G)5(LA-G) + 2Me^a
	473.0868^5−	473.0872^5−	0.9	(L6S-G)5(LA-G) + 3Me^a
	476.8890^5−	476.8893^5−	0.6	(L6S-G)5(LA-G) + 3Me^a
	478.7919^5−	478.7924^5−	1.0	(L6S-G)5(LA-G) + 4Me^a
14	461.7439^6−	461.7441^6−	0.4	(L6S-G)6(LA-G) + 2Me
	464.0798^6−	464.0801^6−	0.7	(L6S-G)6(LA-G) + 3Me
	529.0951^5−	529.0949^5−	0.4	(L6S-G)5(LA-G)2
	531.8975^5−	531.8979^5−	0.8	(L6S-G)5(LA-G)2 + 1Me^a
	534.7004^5−	534.7010^5−	1.1	(L6S-G)5(LA-G)2 + 2Me^a
	537.5040^5−	537.5042^5−	0.4	(L6S-G)5(LA-G)2 + 3Me^a
	540.3076^5−	540.3073^5−	0.6	(L6S-G)5(LA-G)2 + 4Me^a
	543.1106^5−	543.1104^5−	0.4	(L6S-G)5(LA-G)2 + 5Me^a
	466.4160^6−	466.4160^6−	0.0	(L6S-G)6(LA-G) + 4Me
	468.7524^6−	468.7520^6−	0.9	(L6S-G)6(LA-G) + 5Me
16	453.3595^7−	453.3599^7−	0.9	(L6S-G)7(LA-G) + 2Me
	455.3620^7−	455.3621^7−	0.2	(L6S-G)7(LA-G) + 3Me
	457.3643^7−	457.3644^7−	0.2	(L6S-G)7(LA-G) + 4Me
	459.3662^7−	459.3666^7−	0.9	(L6S-G)7(LA-G) + 5Me
	512.7602^6−	512.7600^6−	0.4	(L6S-G)6(LA-G)2 + 2Me
	517.4326^6−	517.4319^6−	1.4	(L6S-G)6(LA-G)2 + 4Me^b
18	447.0708^8−	447.0717^8−	2.0	(L6S-G)8(LA-G) + 2Me
	448.8234^8−	448.8237^8−	0.7	(L6S-G)8(LA-G) + 3Me
	450.5752^8−	450.5757^8−	1.1	(L6S-G)8(LA-G) + 4Me
	501.0915^7−	501.0923^7−	1.6	(L6S-G)7(LA-G)2 + 4Me
	503.0938^7−	503.0945^7−	1.4	(L6S-G)7(LA-G)2 + 5Me
	505.0958^7−	505.0967^7−	1.8	(L6S-G)7(LA-G)2 + 6Me

---

Fig. 3 Negative ESI sequential fragmentations of the porphyran dimethylated tetrasaccharide ion at m/z 426.0750 formed by Pseudoalteromonas atlantica extracts. (A) MS² spectrum of the [dp4 − 2Na]²⁻ precursor ion at m/z 426.0750. (B) MS³ spectrum of the B₂³⁻ ion at m/z 329.0355. (C) MS⁴ spectrum of the B₂ ion at m/z 417.0674. (D) MS⁵ spectrum of the B₁ ion at m/z 241.0022.

The triply charged species at m/z 418.4014 appearing as the most intense ion in the ESI-MS spectrum (Fig. 2) was attributed to a porphyran trisulfated hexasaccharide including two methyl groups. Its MS² fragmentation yielded a major ion at m/z 358.3792 corresponding to a trisulfated pentasaccharide with two methyl groups (Fig. 4A). We assume that it resulted from a glycosidic rupture releasing the terminal galactose unit at the reducing end of the parent hexasaccharide thereby forming the B₅³⁻ ion [M_dp6–G − 3Na]³⁻. Two cross-ring fragments at m/z 388.3895 (^1,4A₆³⁻) and m/z 579.1199 (^2,4A₆²⁻) confirmed the absence of a methyl group on the C-6 of the reducing-end galactose. Nevertheless, an additional B₅³⁻ ion at m/z 353.7074 and ^1,4A₆³⁻ ion at m/z 383.7147 was weakly detected, corresponding to a trisulfated pentasaccharide with only one methyl group (underlined ions in Fig. 4A). This result indicates that a minor proportion of the parent ions had a methylated galactose (G6Me) at the reducing end, thus highlighting the structural heterogeneity of the parent dimethylated trisulfated hexasaccharide with regard to the location of methyl groups. Other glycosidic bond cleavages of the parent hexasaccharide were also detected at m/z 241.0011, m/z 417.0683/435.0785, m/z 329.0346/338.0398, m/z 417.0683/426.0734, m/z 364.3825 for B₁, B₂/C₂, B₃²⁻/C₃²⁻, B₄²⁻/C₄²⁻, C₅³⁻, respectively from the non-reducing end, and at m/z 410.0637/419.0672, m/z 498.0943/507.0994 and m/z 597.1305 for Z₄²⁻/Y₄²⁻, Z₅²⁻/Y₅²⁻ and Y₃ respectively from the reducing end. Further fragmentation by MS³ experiments targeting the B₅³⁻ ion at m/z 358.3792 provided an intense peak at m/z 417.0681 (Fig. 4B) that corresponded to a disulfated tetrasaccharide including two methyl groups. It resulted from the loss of a L6S moiety and was attributed to a B₄²⁻ corresponding to [M_dp6–(L6S-G) − 3Na]²⁻ species. Another minor B₄²⁻ ion at m/z 410.0613 was also observed, corresponding to a disulfated tetrasaccharide with one methyl group [M_dp6–(L6S-G6Me) − 3Na]²⁻. Additional glycosidic bond cleavages were observed, one of them yielding the second-most intense ion (55% of the base peak) at m/z 329.0345 and attributed to a B₃²⁻ fragment, i.e. a disulfated trisaccharide with one methyl group [M_dp6–(G6Me-L6S-G) − 3Na]²⁻. Another B₃²⁻ fragment of lower intensity at m/z 322.0263 was also detected corresponding to [M_dp6–(G6Me-L6S-G6Me) − 3Na]²⁻. Subsequent MS⁴ dissociation experiments performed on the most intense B₄²⁻ at m/z 417.0681 led to two B₃²⁻ fragments (Fig. 4C), an intense one at m/z 329.0355 corresponding to [M_dp6–(G6Me–L6S-G) − 3Na]²⁻, and a minor one at m/z 322.0277 as previously observed in the MS³ step. The MS⁴ spectrum revealed also two cross-ring ions at m/z 308.0307 (^0,2A₃²⁻) and m/z 374.0509 (^1,4A₄²⁻) that confirmed the methylation of the G residue located in the third position from the hexasaccharide reducing end. Further sequential multistage MS experiments were carried out on the most intense ions, continuing up to MS⁷ fragmentation of the non-reducing end terminal residue of the parent hexasaccharide, i.e. a sulfated L-galactose (Fig. 4D). This latter residue was produced upon the MS⁶ fragmentation of the B₂⁻ ion corresponding to the non-reducing disaccharide L6S-G6Me (not shown). The MS⁷ soft fragmentation of B₁ fragment (m/z 241.0015) mainly produced two singly charged ions, at m/z 241.0021 and m/z 96.9602, corresponding to a remaining B₁ precursor ion attributed to non-reducing end monosaccharide L6S and the sulfate anion HSO₄⁻, respectively. Additional ions at m/z 138.9706, 150.9706, 168.9811 and 180.9810 resulting from cross-ring cleavages ^0,4A₁, ^0,3A₁-H₂O, ^0,3A₁ and ^0,4A₁-H₂O respectively, unambiguously positioned the sulfate group at the C-6 of the terminal galactose residue. By combining all these MSⁿ data, we conclude that the parent dimethylated trisulfated hexasaccharide is actually composed of the three following isomers: (L6S-G6Me)–(L6S-G6Me)–(L6S-G), (L6S-G6Me)–(L6S-G)–(L6S-G6Me) and (L6S-G)–(L6S-G6Me)–(L6S-G6Me) (Table 2). Their proportions can be estimated at ≈79, 17 and 4%, respectively, based on the intensity of their fragments and considering similar ionization efficiency of each parent isomer. This ESI-MSⁿ approach thus provides a complete structural elucidation of various isobaric porphyran oligosaccharides.

Fig. 4 Negative ESI sequential fragmentations of a dimethylated hexasaccharide fraction formed by Pseudoalteromonas atlantica extracts. (A) MS² spectrum of the [dp6-3Na]³⁻ precursor ion at m/z 418.4015. (B) MS³ spectrum of the B₅³⁻ ion at m/z 358.3792. (C) MS⁴ spectrum of the B₄²⁻ ion at m/z 417.0681. (D) MS⁷ spectrum of the B₁ ion at m/z 241.0021. Fragments attesting to methylation at the non-reducing end are underlined.

4. Conclusion

The detailed MS sequencing of the degradation products of porphyran incubated with protein extracts from the marine bacterium P. atlantica revealed a new series of oligosaccharides that differed markedly in their level of methylation from those obtained with Z. galactinovorans β-porphyranase A. The methylated disaccharide (L6S-GMe) and dimethylated tetrasaccharide (L6S-GMe–L6S-GMe) end-products formed by the P. atlantica extract have never been observed before and were absent in the end-products of β-porphyranase A. This result suggests that a β-porphyranase with new substrate specificity is present in P. atlantica extracts. The occurrence of a methylated disaccharide end-product demonstrated that P. atlantica β-porphyranase accommodates at least one methyl group in its active site. Furthermore, the observation of a dimethylated tetrasaccharide indicates that methyl group can be positioned at the reducing or non-reducing ends of the oligosaccharide substrate. Therefore, the active site of this novel enzyme, likely a β-methyl-porphyranase, can accommodate methylated porphyran and catalyze a cleavage reaction that cannot be performed by previously described β-porphyranases belonging to the GH16 and GH86 glycoside hydrolase families.^29,38

This study was conducted using protein extracts that may contain several enzymes involved in the degradation of porphyran; among them, one or several enzymes must have methyl porphyran specificities. Previous inspection of the P. atlantica genome revealed that four genes code for a protein belonging to the GH16 family, but one protein Patl_0824 (Q15XN6, http://www.uniprot.org/uniprot/Q15XN6) in particular showed sequence homology with Z. galactinovorans β-porphyranase A.³⁸ This protein has been cloned and overexpressed in E. coli and has similar substrate specificity as Z. galactinovorans β-porphyranases A. Therefore, the β-methyl-porphyranase identified here probably has a different protein sequence from previously described porphyranases or belongs to a new glycoside hydrolase family. The isolation and characterization of this new β-methyl-porphyranase is underway.

Given the hybrid structure of porphyran and its numerous chemical modifications such as methylation, porphyranases with a variety of substrate specificities are required for more in-depth knowledge of the polysaccharide's structure. Although there is a high degree of methylation – above 50% (ref. 45) – at the C-6 position of the D-galactose units, the previously described β-porphyranase cannot catalyze glycosidic bond cleavage within methylated sequences. Therefore, the new methyl β-methyl-porphyranase activity described here will be a valuable structural tool and, combined with β-agarases and β-porphyranases, can produce the building blocks specific to porphyrans from various red algal species.

Acknowledgements

The authors sincerely thank Genopole-France and the Île-de-France Regional Council for generous financial support.

References

1. V. H. Pomin and P. A. S. Mourão, Glycobiology, 2008, 18, 1016.
2. S. Knutsen, D. Myslabodski, B. Larsen and A. I. Usov, Bot. Mar., 1994, 37, 163.
3. S. Peat, J. R. Turvey and D. A. Rees, J. Chem. Soc., 1961, 1590,  10.1039/jr9610001590.
4. S. Peat and D. A. Rees, Biochem. J., 1961, 79, 7.
5. S. Bhatia, A. Sharma, K. Sharma, M. Kavale, B. Chaugule, K. Dhalwal, A. Namdeo and K. Mahadik, Pharmacogn. Rev., 2008, 2, 271.
6. G. A. de Ruiter and B. Rudolph, Trends Food Sci. Technol., 1997, 8, 389.
7. M.-J. Kwon and T.-J. Nam, Life Sci., 2006, 79, 1956.
8. H. Noda, J. Appl. Phycol., 1993, 5, 255.
9. D.-H. Ngo and S.-K. Kim, Int. J. Biol. Macromol., 2013, 62, 70.
10. K. Nisizawa, H. Noda, R. Kikuchi and T. Watanabe, Hydrobiologia, 1987, 151–152, 5.
11. S. Fukuda, H. Saito, S. Nakaji, M. Yamada, N. Ebine, E. Tsushima, E. Oka, K. Kumeta, T. Tsukamoto and S. Tokunaga, Eur. J. Clin. Nutr., 2006, 61, 99.
12. N. Inoue, N. Yamano, K. Sakata, K. Nagao, Y. Hama and T. Yanagita, Biosci., Biotechnol., Biochem., 2009, 73, 447.
13. K. Ishihara, C. Oyamada, R. Matsushima, M. Murata and T. Muraoka, Biosci., Biotechnol., Biochem., 2005, 69, 1824.
14. V. L. Campo, D. F. Kawano, D. B. d. Silva Jr and I. Carvalho, Carbohydr. Polym., 2009, 77, 167.
15. F. van de Velde, Food Hydrocolloids, 2008, 22, 727.
16. A. I. Usov, Food Hydrocolloids, 1998, 12, 301.
17. G. Correc, J.-H. Hehemann, M. Czjzek and W. Helbert, Carbohydr. Polym., 2011, 83, 277.
18. F. van de Velde, S. H. Knutsen, A. I. Usov, H. S. Rollema and A. S. Cerezo, Trends Food Sci. Technol., 2002, 13, 73.
19. M. Lahaye, W. Yaphe, M. T. P. Viet and C. Rochas, Carbohydr. Res., 1989, 190, 249.
20. D. A. Rees and E. Conway, Biochem. J., 1962, 84, 411.
21. Z. A. Popper, G. Michel, C. Hervé, D. S. Domozych, W. G. T. Willats, M. G. Tuohy, B. Kloareg and D. B. Stengel, Annu. Rev. Plant Biol., 2011, 62, 567.
22. N. S. Anderson and D. A. Rees, J. Chem. Soc., 1965, 5880,  10.1039/jr9650005880.
23. D. A. Rees, Biochem. J., 1961, 81, 347.
24. M. D. Noseda, A. G. Viana, M. E. R. Duarte and A. S. Cerezo, Carbohydr. Polym., 2000, 42, 301.
25. M. Duckworth and J. R. Turvey, Biochem. J., 1969, 113, 687.
26. L. M. Morrice, M. W. McLean, W. F. Long and F. B. Williamson, Eur. J. Biochem., 1983, 133, 673.
27. L. M. Morrice, M. W. McLean, W. F. Long and F. B. Williamson, in Eleventh International Seaweed Symposium, eds. C. Bird and M. Ragan, Springer, Netherlands, 1984, vol. 22, ch. 118, pp. 572–575.
28. J. R. Turvey and J. Christison, Biochem. J., 1967, 105, 317.
29. J.-H. Hehemann, G. Correc, F. Thomas, T. Bernard, T. Barbeyron, M. Jam, W. Helbert, G. Michel and M. Czjzek, J. Biol. Chem., 2012, 287, 30571.
30. M. Jam, D. Flament, J. Allouch, P. Potin, L. Thion, B. Kloareg, M. Czjzek, W. Helbert, G. Michel and T. Barbeyron, Biochem. J., 2005, 385, 703.
31. G. Michel, P. Nyval-Collen, T. Barbeyron, M. Czjzek and W. Helbert, Appl. Microbiol. Biotechnol., 2006, 71, 23.
32. W.-J. Chi, Y.-K. Chang and S.-K. Hong, Appl. Microbiol. Biotechnol., 2012, 94, 917.
33. B. L. Cantarel, P. M. Coutinho, C. Rancurel, T. Bernard, V. Lombard and B. Henrissat, Nucleic Acids Res., 2009, 37, D233.
34. J.-H. Hehemann, A. G. Kelly, N. A. Pudlo, E. C. Martens and A. B. Boraston, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 19786.
35. M. Fer, A. Préchoux, A. Leroy, J.-F. Sassi, M. Lahaye, C. Boisset, P. Nyvall-Collén and W. Helbert, J. Microbiol. Methods, 2012, 89, 222.
36. D. Ropartz, A. Giuliani, C. Hervé, A. Geairon, M. Jam, M. Czjzek and H. Rogniaux, Anal. Chem., 2015, 87, 1042.
37. R. Belas, D. Bartlett and M. Silverman, Appl. Environ. Microbiol., 1988, 54, 30.
38. J.-H. Hehemann, G. Correc, T. Barbeyron, W. Helbert, M. Czjzek and G. Michel, Nature, 2010, 464, 908.
39. D. K. Kidby and D. J. Davidson, Anal. Biochem., 1973, 55, 321.
40. C. E. Zobell, J. Mar. Res., 1941, 4, 42.
41. C. Przybylski, F. Gonnet, D. Bonnaffé, Y. Hersant, H. Lortat-Jacob and R. Daniel, Glycobiology, 2010, 20, 224.
42. A. Seffouh, F. Milz, C. Przybylski, C. Laguri, A. Oosterhof, S. Bourcier, R. Sadir, E. Dutkowski, R. Daniel, T. H. van Kuppevelt, T. Dierks, H. Lortat-Jacob and R. R. Vivès, FASEB J., 2013, 27, 2431.
43. D. Ropartz, P.-E. Bodet, C. Przybylski, F. Gonnet, R. Daniel, M. Fer, W. Helbert, D. Bertrand and H. Rogniaux, Rapid Commun. Mass Spectrom., 2011, 25, 2059.
44. B. Domon and C. E. Costello, Glycoconjugate J., 1988, 5, 397.
45. Q. Zhang, H. Qi, T. Zhao, E. Deslandes, N. M. Ismaeli, F. Molloy and A. T. Critchley, Carbohydr. Res., 2005, 340, 2447.

---

Footnotes

† Electronic supplementary information (ESI) available: Table S1 list of oligosaccharides detected by MALDI-TOF MS produced by a Pseudoalteromonas atlantica protein extract; Text S1, Fig. S1 and Table S2, MALDI-TOF MS analysis of porphyran oligosaccharides produced by β-porphyranase A; Text S2 and Fig. S2 sequencing of β-porphyranase A oligosaccharides using negative ESI-MSⁿ. See DOI: 10.1039/c5ra14449c

‡ Present address: Centre de Recherches sur les Macromolécules Végétales, Equipe Chimie et Biotechnologie des Oligosaccharides, UPR CNRS 5301, 38041 Grenoble Cedex, France.

---