Cytochrome b558/566 from the archaeon Sulfolobus acidocaldarius has a unique Asn-linked highly branched hexasaccharide chain containing 6-sulfoquinovose


U. Zähringer, Forschungszentrum Borstel, Zentrum für Medizin und Biowissenschaften, Parkallee 22, D-23845 Borstel, Germany. Fax: +49 4537188612, Tel.: +49 4537188462, E-mail:


Cytochrome b558/566 from the archaeon Sulfolobus acidocaldarius (DSM 639) has been described as a novel highly glycosylated membrane-bound b-type hemoprotein [Hettmann, T., Schmidt, C. L., Anemüller, S., Zähringer, U., Moll, H., Petersen, A. & Schäfer, G. (1998) J. Biol. Chem.273, 12032–12040]. The purified cytochrome b558/566 was characterized by MALDI MS as a 64-kDa (glyco)protein expressing 17% glycosylation. Detailed chemical studies showed that it was exclusively O-mannosylated with monosaccharides and N-glycosylated with at least seven hexasaccharide units having the same unique structure. The hexasaccharide was released by cleavage with peptide:N-glycosidase (PNGase) F and found to consist of two residues each of Man and GlcNAc and one residue each of Glc and 6-deoxy-6-sulfoglucose (6-sulfoquinovose). The last sugar has been known as a component of glycolipids of plants and some prokaryotes, but has not been hitherto found in bacterial glycoproteins. Digestion with trypsin/pronase gave a mixture of glycopeptides with the same Asn-linked hexasaccharide chain, from which an N-glycosylated Tyr-Asn dipeptide was purified by gel chromatography and anion-exchange HPLC. Studies of the degradation products using methylation analysis, ESI MS, MALDI MS, and 1H and 13C NMR spectroscopy, including 1H,13C HMQC and NOESY experiments, established the structure of the unique Asn-linked hexasaccharide chain of cytochrome b558/566.


fraction C glycopeptide


high-performance anion-exchange chromatography




6-sulfoquinovose (6-deoxy-6-sulfoglucose)

Sulfolobus acidocaldarius is an obligate aerobic, hyperthermoacidophilic crenarchaeon growing at 75–80 °C and pH 2–2.5 [1]. In S. acidocaldarius membranes three different a-type, two different b-type but no c-type cytochromes have been detected. All but one cytochrome can be attributed to either the SoxABCD quinol oxidase complex [2,3] or the SoxM terminal oxidase complex [4]. The exceptional cytochrome b558/566, named after its absorption maxima in the redox difference spectrum, demonstrated no similarity in the gene sequence to any entry in data banks [5] and is thus a novel kind of b-type hemoprotein. The isolated cytochrome exhibits a strongly positive redox potential of > 350 mV; indeed, it can be selectively reduced by ascorbate in intact membranes. More interestingly, cytochrome b558/566 expression can be significantly induced by reduced oxygen supply and by amino acids with a branched hydrophobic side chain. Although its physiological function is still enigmatic, it has been inferred that it participates as an ectoenzyme in periplasmatic metabolism [5]. Recently, using an improved purification method, we have isolated cytochrome b558/566 in a high yield and characterized it as a highly glycosylated protein containing O-glycosidically linked mannose residues and at least one N-glycosidically linked oligosaccharide chain [5].

Now we report further structural characterization of cytochrome b558/566 aiming at elucidation of the full structure of the unique Asn-linked oligosaccharide.

Materials and methods

Bacterial growth and isolation of cytochrome b558/566

Cells of S. acidocaldarius (DSM 639) were grown heterotrophically in a mineral salt medium at 78 °C and at pH 2.5 as described [5,6]. Cells were disrupted in a Manton–Gaulin press and membranes prepared as described [6]. After precipitation with ammonium sulfate, cytochrome b558/566 was isolated in pure form by hydrophobic interaction chromatography on propyl-agarose (Sigma P-5268) followed by gel filtration on Highload Superdex 200 (Pharmacia, Sweden) [5].

Enzymatic cleavages

Cytochrome b558/566 (34 mg) in 0.1 m NaHCO3, pH 8.4 (5 mL) was treated with trypsin (1 mg, Boehringer Mannheim, Germany) at 37 °C for 16 h. The mixture was freeze-dried, dissolved in 10 mm sodium phosphate buffer, pH 7.5, containing 0.02% NaN3 (9 mL), and treated with pronase (3.8 mg, Boehringer Mannheim) at 37 °C for 20 h. Treatment with pronase was repeated once and followed by heating at 100 °C for 10 min. Fractionation of water-soluble products on TSK HW-40 (Merck, Germany) (Fig. 1) following desalting on Sephadex G-10 (Pharmacia) gave four glycopeptide fractions: A (0.8 mg), B (3.7 mg), C (1.2 mg), and D (0.3 mg). Fraction C was purified by high-performance anion-exchange chromatography (HPAEC) to give an individual glycopeptide (GP-C, 0.6 mg).

Figure 1.

Electrospray ionization mass spectrum of the hexasaccharide obtained by cleavage of cytochrome b558/566 with PNGase F.

Cytochrome b558/566 (5 mg) in 10 mm sodium phosphate buffer, pH 7.5 (0.4 mL) was treated with peptide:N-glycosidase (PNGase) F (2 units, Boehringer Mannheim) at 37 °C for 4 days. Desalting on Sephadex G-10 followed by gel chromatography on TSK HW-40 yielded a single oligosaccharide (0.1 mg).

Chromatography and GLC-MS

Gel chromatography was carried out on a column of TSK HW-40 in 0.05 m pyridinium acetate buffer, pH 4.5, and monitored with a Knauer differential refractometer (Germany).

HPAEC was performed using a Dionex system (USA) on a CarboPac PA1 column at 1 mL·min−1 in a gradient of NaOAc (0–0.5 m) in 0.1 m NaOH for 60 min. Fractions were desalted on a column of Sephadex G-10 in water.

GLC was performed on a Varian 3700 chromatograph (USA) equipped with a SPB-5 fused-silica gel column (0.25 mm × 30 m) using a temperature gradient from 150 °C (3 min) to 320 °C at 5 °C·min−1 was used. GLC-MS was performed on a HP 5989 A instrument (Hewlett-Packard, USA) equipped with an HP5 column under the same chromatographic conditions as in GLC.

Mass spectrometry

MALDI MS was done with a Bruker-Reflex II instrument (Bruker-Franzen Analytik, Germany) in reflector configuration in the negative linear time-of-flight mode at an acceleration voltage of 20 kV and delayed ion extraction. Samples were dissolved in distilled water at a concentration of 10 µg·µL−1, and 2 µL solution was mixed with 2 µL 0.5 m 2,4,6-trihydroxyacetophenone (Aldrich, USA) in methanol as matrix solution. 0.5 µL aliquots were deposited on a metallic sample holder and analyzed immediately after drying in a stream of air. Mass spectra were recorded in the negative ion mode.

ESI MS was run in negative mode using a VG Quattro triple quadrupole mass spectrometer (Micromass, UK) with aqueous 50% acetonitrile containing 1 mm ammonia as the mobile phase at a flow rate of 10 µL·min−1. Sample was dissolved in aqueous 50% acetonitrile at a concentration of ≈ 50 pmol·µL−1, and 10 µL was injected via a syringe pump into the electrospray source.

NMR spectroscopy

Spectra were run with a Bruker DRX-600 spectrometer (Germany) using a microprobe head at 27 °C in 2H2O. Prior to the measurements, the samples were freeze-dried twice from 2H2O. Standard Bruker software (xwinnmr 1.3) was used to acquire and process the NMR data. Mixing times of 250 and 500 ms were used in TOCSY and NOESY experiments, respectively.

Sugar and methylation analyses

For sugar analysis, samples (0.1 mg) were hydrolyzed with aqueous 2 m CF3CO2H (100 °C, 4 h), reduced with NaBH4 in water, peracetylated with acetanhydride in pyridine (1 : 1.5, v/v, 85 °C, 20 min), and analyzed by GLC [7].

Samples (0.2 mg) were methylated with MeI in Me2SO in the presence of solid NaOH [8]. Partially methylated sugars were derived by hydrolysis with 4 m CF3CO2H (100 °C, 2 h), reduced with NaB2H4, peracetylated, and analyzed by GLC-MS [9].


Cleavage of N-glycosidic linkages in cytochrome b558/566 with PNGase F

HPAEC analysis of the degradation products demonstrated a single oligosaccharide. Sugar analyses indicated that it contained Glc, Man, and GlcNAc in the molar ratios ≈ 1 : 2 : 2. The negative mode ESI mass spectrum (Fig. 1) showed a peak for the pseudomolecular ion [M-H] at m/z 1135.9, indicating a molecular mass higher by 226 Da than that expected for a GlcMan2GlcNAc2 pentasaccharide with the calculated molecular mass of 910.8 Da. The 1H-NMR spectrum of the oligosaccharide contained signals for five glycosidically linked anomeric protons at δ 4.50–5.30 and one sugar residue at the reducing end at δ 5.11 and 4.68 for the α-form and β-form, respectively. These data showed that the isolated product is a hexasaccharide and that one monosaccharide component escaped detection in sugar analysis. Based on the molecular mass difference of 226 Da, it could be suggested that this monosaccharide is a sulfonated deoxyhexose, and further studies revealed that 6-deoxy-6-sulfoglucose (6-sulfoquinovose, Qui6S) is indeed present.

Digestion of cytochrome b558/566 with trypsin/pronase

This treatment resulted in a mixture of glycopeptides that were fractionated by gel chromatography on TSK HW-40 (Fig. 2). Composition, MALDI MS, and NMR spectroscopic analyses showed that fractions A–D are glycopeptides having the same Asn-linked oligosaccharide chain but differing in the peptide sequence. Fraction C was found to be a single glycopeptide (GP-C) that was additionally purified by HPAEC.

Figure 2.

Elution profile of gel chromatography on TSK HW-40 of trypsin/pronase-digested cytochrome b558/566 .

Methylation analysis of GP-C revealed terminal Man and Glc in the ratio 2 : 1, whereas only negligible amounts of partially methylated GlcNAc derivatives were detected. The 1H-NMR spectrum of GP-C (Fig. 3) contained signals for six sugar anomeric protons at δ 4.52–5.29, two N-acetyl groups at δ 2.00 and 2.07, two amino acids, Asn and Tyr, at δ 4.39, 3.62 (both H-α), 2.35–3.09 (H- βa,b), 6.79 and 7.11 (H-2,6 and H-3,5 of Tyr, respectively), and other signals in the region δ 3.15–4.39. The MALDI mass spectrum of GP-C showed a peak of the pseudomolecular ion [M-H] at m/z 1412.8, which was consistent with a GlcQui6SMan2GlcNAc2AsnTyr glycopeptide.

Figure 3.

1H-NMR spectrum and structure of fraction C glycopeptide (GP-C) obtained after trypsin/pronase digestion of cytochrome b558/566 (seeFig. 2).

Elucidation of the structure of GP-C by NMR spectroscopy

The 1H-NMR spectrum of GP-C was assigned using two-dimensional 1H,1H-correlation NMR spectroscopy (Table 1). A TOCSY experiment revealed correlations of H-1 with H-2,3,4,5,6a,6b of both GlcNAc residues (GlcNAcI and GlcNAcII) and with H-2,3,4,5 of Glc and Qui6S, that is, for the four monosaccharides having the gluco configuration. The signals within these spin systems were assigned using COSY and H,H-relayed COSY experiments. These also allowed assignment of H-6a,6b protons of Glc and Qui6S by their correlation to H-5. Spin systems of two GlcNAc residues were distinguished by correlation of the protons at the carbons bearing nitrogen (H-2) to the corresponding carbons (C-2) in the region δ 54–56 revealed by a 1H,13C HMQC experiment. Relatively large coupling constant values, including J1,2-values (9.9 Hz for GlcNAcI and 7.5–8.3 Hz for the others), showed that all gluco-configurated monosaccharides are β-pyranosides. A NOESY experiment confirmed the assignment of their H-3,5 signals, which gave strong cross-peaks with H-1 typical of β-linked pyranosides.

Table 1. 600-MHz 1H-NMR data of fraction C glycopeptide (GP-C) obtained after trypsin/pronase digestion of cytochrome b558/566 (see Fig. 2).
 Chemical shift (p.p.m.)
ResidueH-1H-2 (H-α)H-3 (H-βa)H-4 (H-βb)H-5 (H-2,6)H-6a (H-3,5)H-6bCH3CON
Asn (4.39)(2.87)(3.09)    
Tyr (3.62)(2.35)(2.57)(7.11)(6.79)  

In the TOCSY experiment, H-1 of both Man residues showed correlations to H-2,3 only. However, on the line of the H-2 resonances, there were clear cross-peaks with H-3,4,5, and the signals for H-6b were found by their correlation to H-5 in the COSY spectrum. Coupling constant values showed that both Man residues (ManI and ManII) are in the pyranose form as well. The α configuration of ManI and ManII followed from the absence from the NOESY spectrum of H-1/H-3,5 cross-peaks and the presence of intense H-1/H-2 cross-peaks typical of α-linked pyranosides. Assignment of the signals for Asn and Tyr was achieved by comparison with the data for the model compounds, Asn-Tyr and Tyr-Asn. This demonstrated also the Asn-Tyr peptide sequence in GP-C. As this sequence of amino acids has been only twice found in the polypeptide chain of cytochrome b558/566, two sites of the hexasaccharide attachment can be attributed to N-144 and/or N-164 [5].

The amount of GP-C was insufficient to obtain a 13C-NMR spectrum, but H-detected 1H,13C HMQC experiments showed that, with respect to sugar signals, both 1H NMR and 13C NMR spectra of GP-C and the predominant fraction B glycopeptides (Fig. 2) are practically identical. Therefore, the 13C NMR spectrum of the latter (Fig. 4) was assigned using 1H,13C HMQC and DEPT experiments (Table 2).

Figure 4.

13C-NMR (bottom) and DEPT (top) spectra of fraction B glycopeptides obtained after trypsin/pronase digestion of cytochrome b558/566 (seeFig. 2).

Table 2. 150 MHz 13C NMR data of fraction B glycopeptides obtained after trypsin/pronase digestion of cytochrome b558/566 (see Fig. 2). Owing to a heterogeneity in the peptide moiety, signals of amino acids were too weak to be assigned and their chemical shifts are not given.
 Chemical shifts
  • a–e

    Assignment could be interchanged.

  • f

     Signal is split due to a heterogeneity in the peptide moiety.


The 1H,13C HMQC spectrum showed correlations of the H-6 signals to the corresponding C-6 signals and, thus, confirmed the assignment of the former. Correlation of H-6a,b of Qui6S to C-6 at δ 52.5 demonstrated a 6-deoxy-6-sulfohexose since such a high-field position of the C-6 resonance could only be due to the replacement of OH-6 with an SO3H group (compare published data: δC-6 53.5 for a Qui6S residue in a glycolipid from Rhizobium meliloti 2011 [10]). The H-6a,b chemical shifts of Qui6S, δ 3.15 and 3.61, were also close to published data [10].

The 13C NMR chemical shift data also provided the basis for linkage analysis. The signal for C-1 of GlcNAcI was in a much higher field at δ 79 as compared to the anomeric carbon signals of the other monosaccharides at δ 100.9–104.0. Hence, GlcNAcI is N-glycosidically linked to Asn. The signals for C-4 of Qui6S and GlcNAcI were shifted downfield to δ 82.9 and 80.5, that is, by ≥8 p.p.m. compared with their positions in the spectra of the nonsubstituted sugars [11]. These shifts were caused by glycosylation and indicated that these monosaccharides are 4-substituted. In GlcNAcII, signals for C-3,4,6 were also shifted downfield by 3–5 p.p.m. (to δ 78.0, 76.3 and 67.3, respectively) and, hence, GlcNAcII is at the branching point and carries two side chains. Chemical shifts for C-2,3,4,5,6 of Glc and both Man residues were close to those in free β-Glc and α-Man [11]. This finding is in agreement with the methylation analysis data and confirmed the terminal position of the three hexoses and their anomeric configurations.

Sequence analysis was performed using a NOESY experiment with GP-C, which revealed interresidue correlations between the transglycosidic protons. The following intense cross-peaks were observed: Glc H-1/Qui6S H-4 at δ 4.52/3.54; Qui6S H-1/GlcNAcII H-3 at δ 4.60/4.20; ManI H-1/GlcNAcII H-4 at δ 5.29/3.73; and GlcNAcII H-1/GlcNAcI H-4 at δ 4.59/3.61. These data were consistent with the modes of substitution of these monosaccharides and demonstrated their sequence in the doubly branched oligosaccharide chain. An interresidue cross-peak between ManII H-1 and GlcNAcII H-6 at δ 4.94/3.87 was weak but also conclusive since, as followed from the 13C NMR chemical shift data, position 6 of GlcNAcII was the only possible site of attachment of ManII.

On the basis of the data obtained, it was concluded that the N-linked hexasaccharide chain of cytochrome b558/566 has the following structure:

inline image


In the urkingdom of archaea several glycoproteins, especially surface layer proteins and flagellin proteins, have been studied in detail, including characterization of their sugar moieties [12–14]. Most of these studies were devoted to glycoproteins from the kingdom of euryarchaeota. In this study, it is the first time that the oligosaccharide structure of a crenarchaeotal glycoprotein has been elucidated.

The hexasaccharide structure presented here displays several unusual features. First is the occurence of a trisubstituted GlcNAc residue. Second is the presence of a rare acidic sugar, 6-sulfoquinovose, which has not been hitherto found in any other glycoprotein, although Qui6S is rather common in glycolipids of chloroplasts and photosynthetic bacteria [15–16]. Further studies will have to show if this sulfonated sugar is specific for glycoproteins from Sulfolobus and other Sulfolobales and may consequently be used as a taxonomic marker either for these, or for thermoacidophilic archaea in general. It should be noted that sulfates, but not sugar sulfonates, have been detected in the surface layer proteins of Halobacterium halobium (salinarum) [17]. Third, it is surprising that only one single type of oligosaccharide is N-glycosidically linked to the protein chain, whereas most other glycoproteins studied display a greater variety of oligosaccharide structures. For example, the cell surface glycoprotein of Halobacterium halobium contains an oligosaccharide chain consisting of varying numbers (10–15) of pentasaccharide repeating units, together with several linear tetrasaccharide chains differing in the terminal sugar residue [17]. Another example is a mannan-rich plasma membrane glycoprotein from Themoplasma acidophilum that contains varying amounts of mannose [12–18].

Although the exact number of hexasaccharide chains bound to cytochrome b558/566 remains unknown, the attachment of at least seven hexasaccharide units and 35 single mannose units can be postulated. This suggestion is based on the difference between the molecular mass of the glycosylated protein (64 210 Da) and the gene-derived mass of the polypeptide chain (50 736 Da), together with the Man: GlcNAc: Glc: Qui6S stoichiometry (7 : 2: 1 : 1, respectively) [5]. Two sites of attachment of the hexasaccharide chain, N-144 and/or N-164, could be inferred from the sequence of two amino acid residues still linked to the oligosaccharide after pronase digestion of the glycoprotein as described here. Determination of the other sites of N-glycosylation in cytochrome b558/566 requires isolation and structural characterization of higher glycopeptides and is a subject of further investigations on the detailed structure of this unique glycoprotein.

Based on prediction algorithms [19,20], the major domain of cytochrome b558/566 is a globule built up mainly of β strands and loops. A hydrophobic α-helical domain at the C-terminus serves presumably as the membrane anchor, whereas a second N-terminal hydrophobic motif is not necessarily involved in membrane binding but has the signature of an export signal. These structural elements together support the conjecture of cytochrome b558/566 to function as a membrane residing ectoenzyme. The high degree of glycosylation is considered as protective against the harsh environmental conditions of S. acidocaldarius in its natural habitats. Accordingly, cytochrome b558/566 is most likely exposed to a bulk-phase pH of about 2–2.5; the presence of 35 monosaccharides and up to seven sulfoquinovose-containing hexasaccharide chains would not only be capable of shielding a significant surface area but also to provide a negative surface potential at the locus of the cytochrome due to a low pKa of the sulfonate groups. Moreover, compared to sulfate, sulfonate possesses an improved resistance against hydrolytic cleavage even under the above conditions and the growth temperature of S. acidocaldarius of about 75–85 °C.


This work was supported by the Deutsche Forschungsgemeinschaft (grant Scha 125/17–3 and Scha 125/22–1,2; G.S.) and the Sonderforschungsbereich 470 (project B5; U.Z.). We thank Mr H.-P. Cordes for help with NMR spectroscopy, Dr B. Lindner and Mrs H. Lüthje (Forschungszentrum Borstel, Germany) for MALDI MS analyses, Prof. P.-E. Jansson and Mrs G. Alvelius (Clinical Research Center, Huddinge Hospital, Sweden) for ESI MS analysis, Mrs U. Schombel and Mrs K. Jakob and Mr W. Verheyen for technical assistance.