Applied Microbiology and Biotechnology

, Volume 71, Issue 1, pp 23–33

Bioconversion of red seaweed galactans: a focus on bacterial agarases and carrageenases


  • Gurvan Michel
    • Equipe Glycobiologie Marine, UMR7139 Végétaux Marins et Biomolécules (CNRS/UPMC)
  • Pi Nyval-Collen
    • Equipe Structure des Polysaccharides Marins, UMR7139 Végétaux Marins et Biomolécules (CNRS/UPMC)
  • Tristan Barbeyron
    • Equipe Glycobiologie Marine, UMR7139 Végétaux Marins et Biomolécules (CNRS/UPMC)
  • Mirjam Czjzek
    • Equipe Glycobiologie Marine, UMR7139 Végétaux Marins et Biomolécules (CNRS/UPMC)
    • Equipe Structure des Polysaccharides Marins, UMR7139 Végétaux Marins et Biomolécules (CNRS/UPMC)

DOI: 10.1007/s00253-006-0377-7

Cite this article as:
Michel, G., Nyval-Collen, P., Barbeyron, T. et al. Appl Microbiol Biotechnol (2006) 71: 23. doi:10.1007/s00253-006-0377-7


Agars and carrageenans are 1,3-α-1,4-β-galactans from the cell walls of red algae, substituted by zero (agarose), one (κ-), two (ι-), or three (λ-carrageenan) sulfate groups per disaccharidic monomer. Agars, κ-, and ι-carrageenans auto-associate into crystalline fibers and are well known for their gelling properties, used in a variety of laboratory and industrial applications. These sulfated galactans constitute a crucial carbon source for a number of marine bacteria. These microorganisms secrete glycoside hydrolases specific for these polyanionic, insoluble polysaccharides, agarases and carrageenases. This article reviews the microorganisms involved in the degradation of agars and carrageenans, in their environmental and taxonomic diversity. We also present an overview on the biochemistry of the different families of galactanases. The structure–function relationships of the family GH16 β-agarases and κ-caraggeenases and of the family GH82 ι-carrageenases are discussed in more details. In particular, we examine how the active site topologies of these glycoside hydrolases influence their mode of action in heterogeneous phase. Finally, we discuss the next challenges in the basic and applied field of the galactans of red algae and of their related degrading microorganisms.


Marine macroalgae synthesize a great diversity of polysaccharides, which constitute their cell wall and energy storage. Like land plants, these eukaryotic phyla produce neutral polysaccharides such as starch and cellulose. However, marine algae are characterized by their abundance of sulfated polysaccharides, which have no equivalent in land plants (Kloareg and Quatrano 1988). Red algae (Rhodophyta) produce sulfated galactans, agars and carrageenans, which are the main components of their cell walls (Craigie 1990). These polysaccharides are well known for their gelling properties, used in a variety of laboratory and industrial applications (Mc Hugh 2003). This large family of hydrocolloids is made up of linear chains of galactose, with alternating α(1→3) and β(1→4) linkages. In agarose, the α-linked galactose units are in the L configuration (L unit), whereas they are in D configuration (D unit) in carrageenans (Rees 1969). Agarose defines the unmodified neutral backbone of agars which may hold up to 20% methyl group and sparing sulfate ester group distributed along the polysaccharide chains. Carrageenans are classified according to the number and the position of sulfated ester (S) and by the occurrence of 3,6 anhydro-bridges in the α-linked residues (DA unit) found in gelling carrageenans (Knutsen et al. 1994; Rees 1969). For example, the three most industrially exploited carrageenans, namely, kappa- (κ, DA-G4S), iota- (ι, DA2S-G4S), and lambda- (λ, D2S6S-G2S) carrageenans, are distinguished by the presence of one, two, and three ester-sulfate groups per repeating disaccharide unit, respectively (Fig. 1).
Fig. 1

Chemical structure of the disaccharide units constituting agarose and carrageenan. The μ- and ν-carrageenans are converted in vivo to κ- and ι-carrageenans, respectively, by the action the galactose-6-sulfurylases (EC:

The chemical structures of agars and carrageenans are very heterogeneous and are correlated to the algal sources, the life stages (i.e., gametophyte vs tetrasporophyte), and the extraction procedures of the polysaccharides (Craigie 1990). In the case of carrageenan, this structural complexity is attributed to the occurrence of a mixture of carrageenans in algal extracts as well as the combination of different ideal carrabioses distributed along the polysaccharide chains giving rise to hybrid carrageenan (Bixler 1996; Greer and Yaphe 1984a; Knutsen et al. 1994; Van de Velde et al. 2001). The so-called κ- and ι-carrageenans consequently usually describe galactans composed essentially but not only of κ- and ι-carrabiose motives. The amount and the chemical structure of carrabiose unit variants inserted in κ- and ι-carrageenan chains modulate their gelling properties. One can mention, for example, that the occurrence of biosynthetic precursor mu- (μ, D6S-G4S) and nu (ν, D2S6S-G4S) carrabiose units in κ- and ι-carrageenan chain, respectively, hinders gelification (Bellion et al. 1983; Van de Velde et al. 2002).

Agarose, κ-, and ι-carrageenans form thermoreversible gels in aqueous solutions, the rigidity of which decreases strongly with the degree of sulfation (Fig. 2). Agarose self-assembles into stiff and brittle gels while ι-carrageenan aggregates in very soft and elastic gels. The gelation properties of carrageenans are also strongly dependent on the salts and the ionic strength of the medium. Potassium and calcium chloride are well known to promote the gelation of κ-carrageenan and ι-carrageenan, respectively (Morris et al. 1980; Rees et al. 1982). The fine molecular organizations of these gels as well as the mechanism leading to the aggregation of the macromolecules are still matters of debate. It is now well accepted that gelation is preceded by a disorder (or less ordered)–ordered transition of the macromolecular conformation. Most of the experimental data support a dimeric structure of agarose and carrageenans in the ordered state which means, according to the interpretations of experimental data, either an aggregation into a double helix (intertwined strand) (Anderson et al. 1969; Rees et al. 1982; Viebke et al. 1994) or a duplex of single helices (Bongaerts et al. 1999; Cuppo et al. 2002; Rochas and Rinaudo 1984; Smidsrød and Grasdalen 1982). ι-Carrageenan is the galactan so far most fully characterized in the solid state, namely, as oriented fibers. X-ray diffraction experiments performed on oriented fibers suggest a double-helix conformation of the macromolecule in gels, probably with a parallel packing of the strands (Anderson et al. 1969; Janaswamy and Chandrasekaran 2002). The X-ray diffraction patterns recorded on κ-carrageenan oriented fibers were less resolved than in the case of ι-carrageenan because of the more disordered interchain packing. Nevertheless, on the basis of similarities between the diffraction patterns of the two polymers, it was proposed that the κ-carrageenan could also occur as double helices (Anderson et al. 1969; Bayley 1955; Millane et al. 1988). The case of agarose is the most ambiguous system, as single and double-helix packing have been proposed by crystallographic analysis of oriented fibers of agarose. In the first place, Arnott et al. (1974) have proposed a model of aggregation involving the double helices in the solid state. The re-examination of agarose fibers by Ford and Atkins (1989) has led these authors to propose a single helix model, which seemed to fit the experimental data better. The fairly rigid chains of agarose may indeed not exist as flexible coils in solution and the agarose gelation should consequently involve the direct aggregation of such rigid chains, promoted by solvent effects (Guenet et al. 1993).
Fig. 2

Transmission electron micrographs of negatively stained gel fragments of (a) κ-carrageenan and (b) ι-carrageenan. The gel of κ-carrageenan was promoted by KCl ions, while the aggregation of the ι-carrageenan was induced by CaCl2

The algal polysaccharides constitute a crucial carbon source for a number of marine bacteria. Among those microorganisms, a number of Gammaproteobacteria, Bacteroidetes, and Planctomycetes are key players in the global carbon cycle in the oceans. Those marine bacteria degrade the cell walls of marine algae by secreting specific glycoside hydrolases. The recycling of red algal biomass notably involves microorganisms which produce diverse agarases and carrageenases. As enzymes which naturally attack concentrated gels, galactanases constitute an interesting model to study the heterogeneous phase degradation of sulfated polysaccharides. In combination with liquid chromatography and mass spectrometry techniques, galactanases are also precious tools to analyze the complex structure of carrageenans (Antonopoulos et al. 2005a,b). Moreover, these enzymes are potential powerful biocatalysts to modify agars and carrageenans and alter their properties, producing new, specific algal biomolecules for foods, cosmetics, and pharmaceuticals (De Ruiter and Rudolph 1997).

In this review, we first examine the microorganisms involved in the degradation of agars and carrageenans, in their environmental and taxonomic diversity. An overview is also presented on the biochemistry of galactanases and on the different structural families of sequenced agarases and carrageenases. At the light of the recent crystal structures of several galactanases, the structure–function relationships of the family GH16 β-agarases and κ-caraggeenases and of the family GH82 ι-carrageenases are discussed in more details. We notably look at the relationships between the active site topology of galactanases and their mode of action in heterogeneous phase. Finally, we discuss the next challenges in the basic and applied field of red seaweed galactans and of their related degrading microorganisms.

Marine bacteria: key microorganisms in the bioconversion of agars and carrageenans

The bioconversion of sulfated galactans from red algae is essentially performed by marine bacteria. No marine fungus has been found yet that displays such activities. Galactanolytic activities have been reported in marine mollusks (Usov and Miroshnikova 1975), but recent studies on marine herbivores, such as abalones (Erasmus et al. 1997) and fishes (Skea et al. 2005), have demonstrated that the degradation of galactans is mainly carried out by resident gut bacteria.

The agarases

The first agarolytic bacterium was isolated from seawater at the beginning of the 20th century by Gran (1902). Since then, a number of microorganisms have been reported to degrade agars, mainly in marine environment, either in the water column, in coastal marine sediments or associated to red algae (Humm 1946; Stanier 1941). Agarolytic bacteria have also been identified in brackish water and salt marshes (Ekborg et al. 2005) and, more surprisingly, in fresh water (Van der Meulen et al. 1974) and soils (Buttner et al. 1987; Suzuki et al. 2003). This unexpected presence of agarases in non-marine environments was also confirmed by metagenomic approach on soil samples (Voget et al. 2003). Despite the number of isolated species, the agarolytic bacteria represent only a few phyla and classes. Those marine microorganisms mainly belong to the Gammaproteobacteria class of the Proteobacteria phylum, including the genera Pseudomonas (Ha et al. 1997; Vattuone et al. 1975), Alteromonas (Kimura et al. 1999; Leon et al. 1992; Potin et al. 1993; Young et al. 1971), Pseudoalteromonas (Schroeder et al. 2003; Vera et al. 1998; Yaphe 1957), Vibrio (Aoki et al. 1990; Araki et al. 1998; Sugano et al. 1993b) Alterococcus (Shieh and Jean 1998), Microbulbifer (Ohta et al. 2004ac), Agarivorans (Ohta et al. 2005a), Thalassomonas (Ohta et al. 2005b), and Saccharophagus (Ekborg et al. 2005). Agarolytic bacteria are also found in the three classes of the Bacteroidetes phylum, the Bacteroides class including the genus Marinilabilia (Veldkamp 1961), the Flavobacteria class including the genera Flavobacterium (Van der Meulen et al. 1974), Cellulophaga (Lewin 1969), and Zobellia (Barbeyron et al. 2001), and the Sphingobacteria class including the genera Cytophaga (Turvey and Christison 1967; Veldkamp 1961), Persicobacter (Stanier 1941), and Microscilla (Zhong et al. 2001). In soils, the agarolytic microorganisms are gram-positive bacteria belonging to the Actinobacteria (Stanier 1942) and Firmicutes (Suzuki et al. 2003) phyla.

With the exception of the enzyme from Alteromonas agaralytica (Potin et al. 1993), all the known agarases hydrolyze the β-(1→4) linkage of agarose, yielding oligosaccharides in the series related to neoagarobiose [O-3,6-anhydro-α-l-galactopyranosyl-(1→3)-d-galactose]. The best-characterized agarolytic system has been described in Pseudoalteromonas atlantica (Day and Yaphe 1975; Groleau and Yaphe 1977). The extracellular endo-β-agarase I depolymerizes agarose to neoagarotetraose. Its end-products are further hydrolyzed by two periplasmic enzymes, β-agarase II (also known as neoagarotetraose hydrolase) and neoagarobiose hydrolase, finally yielding 3,6-anhydro-l-galactose and d-galactose for use as a carbon source (Morrice et al. 1983a,b). A similar mode of agar degradation was reported for Pseudomonas elongata (Vattuone et al. 1975), Flavobacterium flevense (Van der Meulen et al. 1974), and Zobellia galactanivorans (Jam et al. 2005). In constrast, Vibrio sp. JT0107 secretes two β-agarases that depolymerize agarose to neoagarobiose and neoagarotetraose, respectively (Sugano et al. 1993a, 1994b), which are subsequently degraded by a periplasmic α-l-galactosidase that cleaves the α-(1→3) linkages of the neoagaro-oligosaccharides from the non-reducing ends (Sugano et al. 1994a).

Eleven β-agarases, with a demonstrated biochemical function, have been sequenced (Belas 1989; Buttner et al. 1987; Jam et al. 2005; Ohta et al. 2004ac; Schroeder et al. 2003; Sugano et al. 1993a, 1994b). The 101-kb plasmid of a bacteria agar-degrading isolate, Microscilla sp., harbors five genes with significant sequence similarities with known β-agarases. Although the individual function of these genes was not experimentally confirmed, the loss of this plasmid renders Microscilla sp. strain PRE1 unable to degrade agarose (Zhong et al. 2001). In their metagenomic study of soil samples, Voget et al. (2003) have also identified three different putative β-agarases genes, referred to as aguA, aguB, and aguF. Overexpression experiments in Escherichia coli demonstrated that aguB and aguF indeed encode active β-agarases. Although all these proteins hydrolyze the β-(1→4) linkage in agarose, sequence analyses indicate that they belong to three unrelated families of glycoside hydrolases, the families GH16, GH50, and GH86 (Henrissat and Bairoch 1996). Table 1 summarizes the current data on these enzymes. The β-agarases from family GH16 are the most characterized mechanistically as well as structurally (see below). Finally, we have recently cloned and sequenced the gene of the α-agarase of A. agaralytica. This enzyme, which cleaves the α-(1→3) linkage of agarose (Potin et al. 1993), defines a new structural family of glycoside hydrolases, the family GH96 (Flament et al., in preparation,
Table 1

Census of the sequenced bacterial agarases and carrageenases



Genpept code


PDB code

Family GH16

 Pseudoalteromonas gracilis B9

β-Agarase AagA


Schroeder et al. 2003


 Microbulbifer sp. JAMB-A7

β-Agarase AgaA7


Ohta et al. 2004b


 Microbulbifer sp. JAMB-A94

β-Agarase AgaA


Ohta et al. 2004c


 Microscilla sp. PRE1

β-Agarase MS116


Zhong et al. 2001


β-Agarase MS132


Zhong et al. 2001


 Pseudoalteromonas carrageenovora

κ-Carrageenase CgkA

AAW 20552

Barbeyron et al. 1994


 Streptomyces coelicolor A3(2)

β-Agarase DagA


Buttner et al. 1987


 Uncultured bacterium

β-Agarase AguB


Voget et al. 2003


β-Agarase AguF


Voget et al. 2003


 Zobellia galactanivorans

β-Agarase AgaA


Jam et al. 2005

1O4Y, 1URX


β-Agarase AgaB


Jam et al. 2005



κ-Carrageenase CgkA


Barbeyron et al. 1998


Family GH50

 Agarivorans sp. JAMB-A11

β-Agarase AgaA11


Ohta et al. 2005a


 Uncultured bacterium

β-Agarase AguA


Voget et al. 2003


 Vibrio sp. JT0107

β-Agarase AgaA


Sugano et al. 1993a


β-Agarase AgaB


Sugano et al. 1994b


Family GH82

 Alteromonas fortis

ι-Carrageenase CgiA


Barbeyron et al. 2000

1H80, 1KTW

 Zobellia galactanivorans

ι-Carrageenase CgiA


Barbeyron et al. 2000


Family GH86

 Microbulbifer sp. JAMB-A94

β-Agarase AgaO


Ohta et al. 2004a


 Microscilla sp. PRE1

β-Agarase MS109


Zhong et al. 2001


β-Agarase MS115


Zhong et al. 2001


β-Agarase MS130


Zhong et al. 2001


 Pseudoalteromonas atlantica T6c

β-Agarase AgrA


Belas 1989


Family GH96

 Alteromonas agarilytica GJ1B

α-Agarase AgaA


Flament et al., in preparation


The carrageenases

In comparison to agar-degrading bacteria, much fewer microorganisms have been reported to hydrolyze carrageenans. All these bacteria were isolated in marine environment and belong to Gammaproteobacteria, Flavobacteria, or Sphingobacteria classes. The first enzyme found to modify carrageenans was described in Japan by Mori (1943). Twelve years later, Yaphe and coworkers isolated nine carrageenolytic Gammaproteobacteria in North-Atlantic Ocean, referred to as strains 1–9, but only strains 1 and 9 were subjected to detailed studies (Bellion et al. 1982; Yaphe and Baxter 1955). The strain 9, which was identified as Pseudoalteromonas carrageenovora (Gauthier et al. 1995), displays two different carrageenase activities, specific for κ-carrageenan and λ-carrageenan, respectively (Weigl and Yaphe 1966). Strain 1, referred to as Alteromonas fortis (Potin 1992), produces one extracellular carrageenase specific for ι-carrageenan (Greer and Yaphe 1984b). A κ-carrageenase has also been purified from a Cytophaga bacterium, referred to as strain 1 k-C783 (Sarwar et al. 1987). Z. galactanivorans, a Flavobacteria isolated from the red alga Delesseria sanguinea in Roscoff (Potin et al. 1991), secretes one κ-carrageenase (Potin et al. 1991) and one ι-carrageenase (Barbeyron et al. 2000), together with two β-agarases (Jam et al. 2005).

All the known carrageenases specifically cleaves the β-(1→4) linkage of their respective substrates. P. carrageenovora is the most studied bacterium that degrades carrageenans. Its κ-carrageenase has been purified (32 kDa) and showed to yield κ-neocarratetraose-sulfate and κ-neocarrabiose-sulfate as end-products (McLean and Williamson 1979b; Weigl and Yaphe 1966). This enzyme proceeds according to a mechanism retaining the anomeric configuration (Potin et al. 1995). κ-Neocarratetraoses are further degraded by a κ-neocarratetraose hydrolase (McLean and Williamson 1981) and a specific sulfatase subsequently hydrolyzes the 4-sulfate group of κ-neocarrabioses (McLean and Williamson 1979a). The λ-carrageenase activity of P. carrageenovora was proposed to be due to an extracellular enzyme complex involving three hydrolases (Johnston and McCandless 1973); but, a single extracellular λ-carrageenase of 98 kDa was purified and showed to hydrolyze lambda family carrageenans, λ-, ξ-, and π-carrageenans (Greer 1984; Potin 1992). The ι-carrageenases from A. fortis and Z. galactanivorans have similar molecular weight (54 kDa) and both release ι-neocarratetraose-sulfate and ι-neocarrahexaose-sulfate as end-products (Barbeyron et al. 2000; Greer and Yaphe 1984b). In contrast to κ-carrageenases, ι-carrageenases proceed according to a mechanism inverting the anomeric configuration (Barbeyron et al. 2000).

Only four genes of carrageenases have been cloned, the κ-carrageenase gene cgkA from P. carrageenovora (Barbeyron et al. 1994), the ι-carrageenase gene cgiA from A. fortis (Barbeyron et al. 2000), and the κ- (cgkA) and ι-carrageenase (cgiA) genes of Z. galactanivorans (Barbeyron et al. 1998, 2000). The κ-carrageenases belongs to the family GH16, which also contains several β-agarases. In contrast, the ι-carrageenases constitute a distinct family of glycoside hydrolases, the family GH82 (Table 1).

β-agarases and κ-carrageenases from family GH-16

All structurally characterized enzymes to date that degrade agarose or κ-carrageenan belong to the same fold family GH-16. This glycoside hydrolase family indeed contains enzymes that proceed via a retaining mechanism, covering a broad variety of substrate specificities, such as lichenases (EC, xyloglucan endotransferases (XET) (EC, keratan-sulfate endo-1,4-beta-galactosidases (EC, glucan endo-1,3-beta-d-glucosidases (EC, endo-1,3(4)-beta-glucanases (EC, xyloglucanases (EC, β-agarases (EC, and κ-carrageenases (EC (see Structures have been reported for several endo-1,3(4)-beta-glucanases (Hahn et al. 1995; Keitel et al. 1993), one XET enzyme (Johansson et al. 2004), two β-agarases (Allouch et al. 2003), one β-galactosidase (Tempel et al. 2005) and one κ- carrageenase (Michel et al. 2001a). Both β-agarases AgaA and AgaB from Z. galactanivorans as well as the κ-carrageenase CgkA from P. carrageenovora adopt a similar jellyroll fold, formed of two seven-stranded β-sheets interconnected by extended loops. The concave β-sheet forms a long active site cleft that runs across the surface of the enzymes, surrounded by extended loops that are determinants not only for differences in specificity but also for the mode of action (see below). Glycosidic bond cleavage by retaining enzymes takes place via a double-displacement mechanism that involves two strictly conserved carboxyl groups (Koshland 1953; Sinnott 1990). These catalytic residues are located in the center of the inner, concave β-sheet, constituted by the characteristic sequence pattern ExDx(x)E. The two glutamates, first and last residues of this pattern, have unambiguously been identified in the B. macerans 1,3-1,4-β-glucanase (Hahn et al. 1995) as nucleophile and acid/base residues, respectively. It was suggested that the role of the third, strictly conserved residue (the central aspartate residue) was to maintain the nucleophilic glutamate in the negatively charged state (Kleywegt et al. 1997). In addition, it is believed that D165 in the κ-carrageenase facilitates the decay of the final transition state in the catalytic cycle (Michel et al. 2001a).

The structural details of the catalytic machinery of the different enzyme structures within family GH-16 are highly similar and local residue substitutions are responsible for the fine tuning of substrate specificities at the active site. The largest differences, however, are encountered in the extended loops that surround the central active site groove, which will, therefore, play a determinant role in modulating the overall topology of the active site entrance. There is general agreement that the topology of enzymes that degrade carbohydrate polymers fall into only three general classes (Davies and Henrissat 1995), which are the pocket or crater, the cleft or groove, and the tunnel topology. It is further accepted that these topologies are indicative of the mode of action of the respective enzymes in that the pocket is encountered in enzymes that degrade the substrate from its extremity such as exo-polysaccharidases, the groove is ideal for random binding of several sugar units in polymeric substrates and is commonly found in endo-acting enzymes, while the tunnel topology will occur in enzymes that display a processive manner of polymer degradation. It is interesting that the different structures of enzymes from family GH-16 cover all these different topologies (Fig. 3): the recent crystal structure of a β-galactosidase from Clostridium perfringens (Tempel et al. 2005) revealed a pocket-formed catalytic active site, while both β-agarases feature an open-groove topology, and the κ-carrageenase CgkA displays a tunnel topology, reminiscent of processive hydrolases (Divne et al. 1994). The presence of all topologies within the same glycoside hydrolase (mechanistic) family makes it an interesting group of enzymes to address evolutionary questions that have already been evoked by Michel et al. (2001a).
Fig. 3

Active site topologies in family GH16 enzymes. β-Agarases AgaA (a) and AgaB (b) from Z. galactanivorans (cleft). κ-Carrageenase from P. carrageenovora (tunnel, c). endo-beta-Galactosidase from C. perfringens (pocket, d)

These structural data have been further confirmed by experimental techniques that monitor the enzymatic degradation of agarose and κ-carrageenan such as HPAEC (high-performance anion exchange chromatography) (Jam et al. 2005) and SEC (size exclusion chromatography) (Nyvall-Collen, in preparation), respectively. On molten agarose, both β-agarases produced high-molecular mass oligosaccharides in the first place, which were then progressively converted into smaller oligosaccharides, a product pattern typical for endo-acting enzymes. Both β-agarases degraded ∼100% of the substrate at completion (24 h) when confronted with liquid-phase agarose. In the solid phase, however, the two enzymes differ markedly in their degradation capacities. Based on the reducing-sugar assay, AgaA degrades at completion 42% of a 0.125% (w/v) agarose gel, whereas AgaB degrades only 21% of the solid gel (Jam et al. 2005). This difference in catalytic behavior towards the gelified substrate possibly is explained by the presence of a surface agarose-binding site in AgaA, revealed by the crystal structure, and that has no equivalent in AgaB (Allouch et al. 2004). The presence of a surface binding site has also been proposed by the analysis of the crystal structure β-helical ι-carrageenase (see below) (Michel et al. 2003). These additional parallel binding sites are thought to either unwind the agarose double helices (Allouch et al. 2004) by acting as a wedge or disrupt the crystalline aggregates of helices constituting the fibers (Michel et al. 2003).

Mode of action of family GH82 ι-carrageenases in heterogeneous phase

The structural fold of family GH82 enzymes revealed a right-handed β-helix, as determined for the ι-carrageenase CgiA from A. fortis (Michel et al. 2001b), a fold that has first been encountered for pectate lyase C (Yoder et al. 1993) but that is also found in polygalacturonases from family GH28 (Petersen et al. 1997). Many of these highly diverse enzymes notably share the common property to interact with and to depolymerize highly acidic polysaccharides present in the extracellular matrix of plants or animals (Michel et al. 2001b). The main difference between the ι-carrageenase and other enzymes displaying the β-helix fold is the presence of two supplementary domains in the C-terminal region, one of which has been shown to undergo large conformational changes upon binding of substrate molecules. This highly flexible module also displays significant similarity to polyanion-binding modules such as RNA/DNA binding modules, and possibly represents a new type of carbohydrate binding module (CBM), yet specific for binding to ι-carrageenan. Ionic interactions between conserved basic protein residues and the sulfate substituents of the polysaccharide chain dominate ι-carrageenan recognition. Glu245 and Asp247 are the proton donor and the base catalyst, respectively (Michel et al. 2003).

The degradation mechanism of ι-carrageenan gels by the ι-carrageenase CgiA from A. fortis has been investigated by transmission electron microscopy, monitoring the degree of ι-carrageenan gel digestion by the release of reducing ends in parallel. At the stages of 20% or even 35% of gel digestion, the initial structure of the fiber network was still apparent, whereas the size of the fibers was deeply modified. Indeed, the lengths of individual fibers decreased in subsequent steps and the thinning of crystalline fibers was strongly amplified after each step of digestion. As in the case of cellobiohydrolases and chitanases, this fiber thinning is indicative of a highly processive mode of action (Michel et al. 2003). The molecular bases of this processivity were established by crystallographic studies of the ι-carrageenase CgiA in complex with ι-carrageenan oligosaccharides. In the presence of substrate, the flexible module A shifts towards the β-helix groove, forming a tunnel which encloses a ι-carrageenan tetrasaccharide, while the N-terminal region binds a disaccharide. Thus, from an open conformation, which facilitates an initial endo-attack of ι-carrageenan chains, the enzyme switches to a closed tunnel form, consistent with its highly processive character. The authors also proposed that the markedly basic outer surface of the enzyme, opposite to the active site tunnel, might displace calcium ions present in ι-carrageenan gel-like aggregates and act as a splitting wedge that threads double helices apart from the ι-carrageenan fibers (Michel et al. 2003). Such a polycationic surface binding site has been confirmed in the structure of the β-helical chondroïtinase B in complex with oligosaccharides of dermatan-sulfate (Michel et al. 2004).

Conclusion: next challenges in the bioconversion of sulfated galactans

The enzymatic degradation of solid polysaccharide biomass remains one of the key reactions in maintenance of the biosphere. It is catalyzed by numerous synergistically acting partners that range from exolytical enzymes or endopolysaccharidase, active on soluble oligosaccharides or single chains, to processive enzymes or those active on solid substrate. While the details of the mechanisms by which bond cleavage is performed by these enzymes is becoming increasingly well understood, the mode of action of those degrading polymeric insoluble substrate remains a challenging and poorly understood area. Furthermore, the diversity and the degree of sophistication of the enzymes involved in the degradation of polysaccharide gels clearly point to the complexity of the substrate. In parallel, the recent advances in the structural analysis of polysaccharide gels (in vitro) give us an insight into the high degree of structural and chemical complexity of the elements that constitute these polysaccharide gels. It will be obviously necessary to join the efforts in understanding the mode of action of these enzymes and the structural and chemical complexity of the solid substrate that is degraded to get an apprehension of the complete picture. The use of specific enzymes will help dissect the structure and forces forming the solid substrate and the detailed knowledge of the substrate complexity will give indications on how the specificity of the enzymes is modulated.

The processive mode of action of numerous polysaccharide-degrading enzymes is a generally accepted feature; although it is not yet completely understood by which molecular mechanism it is fulfilled. Extensive studies on cellobiohydrolase, the first described processive enzyme, have emerged the image that extensive conformational changes of loops opening and closing an active site tunnel on the one hand, and hydrophobic planes on which the substrate may slide on the other hand are key features for the processive advancement of neutral uncharged substrate polymers (Varrot et al. 2003). In contrast, the involvement of conformational changes of positively charged residues in the processivity on acidic substrate has been demonstrated for chitinases (Sorbotten et al. 2005). Crystallographic studies have established a number of structural bases for processivity. The cellobiohydrolases of families GH-6, GH-7, and GH-48 display different folds, but they all exhibit an active site with a tunnel topology (Divne et al. 1994; Parsiegla et al. 1998; Rouvinen et al. 1990). This active site topology allows the protein to progress unidirectionally along the polysaccharide chain without dissociating from its substrate between two hydrolysis events. The presence of an additional carbohydrate-binding module, which remains loosely associated to the crystalline substrate, can also promote the processive behavior of some enzymes with an open-groove topology. Such a CBM was observed in the endo/exo cellulase E4 (Sakon et al. 1997).

The mechanisms established for the degradation of insoluble, neutral polysaccharides cannot be easily transposed to the processive degradation of negatively charged galactans. Both κ- and ι-carrageenases display tunnel-shaped active reminiscent of the cellobiohydrolases. These tunnels explain how the carrageenases remain attached to their substrate after the initial attack. The sliding mechanism on such polyanionic polymers is not resolved yet, but the nature of the substrate recognition clearly indicates that the sliding do not involve hydrophobic platforms as seen in cellobiohydrolases (Varrot et al. 2003). The comparison with processive enzymes acting on nucleic acids is more likely to give answers. Even though the exact nature of the dimeric structure of agarose and carrageenans is still unsolved (double helices vs dimers of single helices), the processive degradation of galactan gels involves the disruption of the polysaccharide fibers. The recent structural analysis of β-agarase AgaA (Allouch et al. 2004) and ι-carrageenase CgiA (Michel et al. 2003) in complex with their respective substrate oligosaccharides possibly point to a novel mode of processivity. In these enzymes a supplemental surface binding site (Fig. 4), together with the binding site in the catalytic groove turns the protein into a sort of wedge that progressively separates substrate aggregates before the catalytic cleavage, performed by the catalytic machinery.
Fig. 4

Structural bases for the processivity of bacterial agarases and carrageenases. Surface agarose binding site in AgaA (a). Tunnel-shaped active site in κ-carrageenase (b) and ι-carrageenase (c). The surfaces colored red correspond to the catalytic residues. The blue surfaces correspond to the putative basic ι-carrageenan binding site in the outer surface of ι-carrageenase

The global understanding of the sulfated galactan bioconversion is still in its infancy. Structural and mechanistic data are lacking for the β-agarases of families GH50 and GH86, for family GH96 α-agarases and for λ-carrageenases. Even in well-characterized families, paralogous enzymes, such as AgaA and AgaB (Jam et al. 2005), may differ in their mode of action. Moreover, the glycoside hydrolases are only one of the classes of enzymes acting on these complex galactans. We know almost nothing about sulfatases or carbohydrate esterases involved in the degradation of agars and carrageenans. One exciting perspective is the emergence of genomic data on galactanolytic marine bacteria. In 2003, Rudolph Amann and coworkers (Max Planck Institute, Bremen) have published the complete genome sequence of Rhodopirellula baltica, a marine bacterium isolated from marine snows (Glöckner et al. 2003). This bacterium from the Planctomycetes phylum features one putative κ-carrageenase gene and, more surprisingly, about 100 genes coding formylglycine-dependent sulfatases. Preliminary data from our group confirm that R. baltica indeed degrades κ-carrageenan, suggesting that this marine bacterium is a good model to study the degradation of sulfated polysaccharides from algae. The draft genome of Saccharophagus degradans has also revealed the presence of several β-agarases of different structural families (Weiner and collaborators, DOE Joint Genome Institute). In collaboration with the French sequencing center, Génoscope, and our German colleagues from the MPI Bremen, we have also started the sequencing of the complete genome of Z. galactanivorans. In combination with classical biochemical approaches, these genomic data should lead to the emergence of model bacteria for dissecting the full pathways for the bioconversion of red seaweed galactans.

Copyright information

© Springer-Verlag 2006