**3.1 GDP-ManA biosynthesis pathway**

It is known that bacterial alginate is initially synthesized as a linear poly-Dmannuronic acid homopolysaccharide before undergoing chemical modifications and epimerization in the periplasmic space, which implies in mannuronic acid (ManA) being the main precursor molecule specifically involved in the biosynthesis of bacterial alginate. However, prior to its utilization to form the poly-D-ManA chain that initiates alginate biosynthesis, ManA must first be synthesized from fructose-6-phosphate (F6P), which is derived from the cell's central carbon metabolism, and then activated with a high-energy bond to yield the sugar nucleotide GDP-ManA [109]. All steps leading to the conversion of F6P to GDP-ManA, the basic building block of alginate, occur in the cytosol and have been extensively analyzed and reported in the past, and a summary of this process is shown in **Figure 3**.

Three enzymes are directly involved in GDP-ManA biosynthesis, namely, AlgA (isomerase/pyrophosphorylase, AlgC (phosphomutase), and AlgD (dehydrogenase). GDP-ManA biosynthesis begins with the conversion of F6P to mannose-6-phosphate (M6P). This reaction is catalyzed by the bifunctional enzyme AlgA through its phosphomannose isomerase (PMI) activity [110]. Through the formation of M6P,

#### **Figure 3.**

*Biosynthesis of guanosine diphosphate-mannuronic acid (GDP-ManA), the precursor molecule for bacterial alginate biosynthesis. AlgA: bifunctional phosphomannose isomerase/GDP-Mannose pyrophosphorylase; AlgC: phosphomannomutase; AlgD: GDP-mannose dehydrogenase; PMI: phosphomannose isomerase activity; GMP: guanosine-diphosphomannose pyrophosphorylase activity; F6P: D-fructose-6-phosphate; M6P: D-mannose-6-phosphate; M1P: D-mannose-1-phosphate; GDP-Man: guanosine diphosphate D-mannose; GDP-ManA: guanosine diphosphate D-mannuronic acid.*

F6P-derived carbon molecules are committed to the alginate biosynthetic pathway, and M6P is rapidly converted to mannose-1-phosphate (M1P) by the Mg2+-dependent phosphomannomutase AlgC [111], a crucial enzyme not only for the biosynthesis of alginate but also required for the production other exopolysaccharides [112], lipopolysaccharides, and rhamnolipids [113, 114]. AlgA is required once again in the next reaction step, however, it will now employ its guanosine-diphosphomannose pyrophosphorylase (GMP) activity to convert M1P to GDP-mannose, simultaneously hydrolyzing GTP and pyrophosphate (PPi). GDP-ManA is finally obtained by the irreversible oxidation of GDP-mannose by AlgD, an NAD<sup>+</sup> -dependent guanosinediphospho-D-mannose dehydrogenase that catalyzes the limiting step and major control point of alginate biosynthesis [41, 115, 116].

#### **3.2 Biosynthesis of the poly-D-ManA linear chain**

Polymerization of GDP-ManA precursors to polymannuronate is carried out by two enzymes located at the inner membrane of bacteria, interfacing the cytosolic material with the other periplasmatic proteins in the alginate biosynthetic complex. Both enzymes, namely Alg8 and Alg44, are indispensable for alginate production [65, 69, 117]. Alg8 is believed to be the catalytic subunit of alginate polymerase since it shares homologies with class II glycosyltransferases [118], which are responsible for the transfer of sugar units from an activated monosaccharide donor to an acceptor molecule. Similarities in the polymerization mechanism of alginate with the biosynthesis of other polysaccharides have been suggested [119], and since the catalytic mechanism of class II glycosyltransferases requires an inversion of the anomeric configuration at the reaction center, in order to yield β-linked products such as the nascent alginate polymer (poly-β-D-ManA) the donor substrates must be and α-linked sugar nucleotide (i.e.: GDP-α-D-ManA) [69] (**Figure 4**). Membrane topology analysis of Alg8 has shown that this protein has at least four transmembrane helices, a short periplasmic loop, and an extended cytoplasmic loop containing the glycosyltransferase domain [118–120]. It has been reported that *alg8* overexpression in both *A. vinelandii* and *P. aeruginosa* greatly increases alginate production in comparison to wild-type strains [80]. The molecular weight of alginate polymers has also been shown to be higher when *alg8* expression increases in *A. vinelandii* [121].

The copolymerase Alg44 is the most controversial protein in the entire alginate biosynthesis complex. It is mostly accepted as being a transmembrane protein localized in the bacterial inner membrane and bears a c-di-GMP-binding PilZ domain at its cytosolic N-terminal portion [79, 103, 104]. However, because its complete structure has not yet been determined, and due to its homology with MexA, a periplasmic membrane fusion protein involved in the multi-drug efflux system [122], some reports have proposed that Alg44 lacks transmembrane domains and instead participates in the multi-protein scaffold formation of the alginate biosynthetic complex bridging Alg8 in the inner membrane to the porin (AlgE or AlgJ) in the outer membrane [79, 112, 117]. As previously stated, however, Alg44 is required for the polymerization of ManA together with Alg8, and its regulatory function via the second messenger c-di-GMP in the Alg8-Alg44 complex has been demonstrated [45, 123]. One particular c-di-GMP synthesizing protein, namely MucR, has been shown to specifically influence the alginate biosynthesis in both *P. aeruginosa* [124] and *A. vinelandii* [125], and mutations in the PilZ domain of Alg44 has resulted in the loss of alginate production [112].

**Figure 4.**

*Proposed mechanism reaction for the polymer-level epimerization of β-D-ManA into α-L-GulA by the periplasmic C5-mannuronan epimerase AlgG.*

#### **3.3 Periplasmic chemical modifications of the polymannuronic acid**

After polymerization by the Alg8-Alg44 complex, the nascent poly-D-ManA chain is directed across a multi-enzymatic scaffold structure situated in the periplasmic space, where ManA units may suffer epimerization or O-acetylation [10, 65, 80]. The enzymes involved in these processes are the best-analyzed proteins in the biosynthesis process of alginate [126, 127], and the understanding of their reaction mechanisms is an important factor to enable the production of tailored polymers [10, 128]. Although neither epimerization nor O-acetyl-esterification is essential for the bacterial production of alginate, these chemical modifications significantly alter the physicochemical properties of the polymer [31, 32, 38].

The O-acetylation of alginate is a process naturally occurring only in bacteria. During its transit through the periplasm, some ManA units of the polymannuronate chain are O-acetyl-esterified at positions C-2 and/or C-3 by the combined activity of four enzymes: AlgX, AlgF, AlgI, and AlgJ/AlgV [10, 65, 69, 77, 110]. The acetyl groups from the donor molecules (probably acetyl-CoA, but still undetermined), are transported from the cytosol to the periplasm by AlgI, a 7-helical transmembrane protein with homology to DltB (D-alanyltransferase) in *Bacillus subtilis* [76], to AlgJ (in *P. aeruginosa*) or AlgV (in *A. vinelandii*). AlgJ/AlgV is found on the periplasm, closely associated with AlgI, and also anchored in the cytoplasmic membrane presumably by its hydrophobic signal peptide [129]. It shows high homology with AlgX, another protein present in the alginate biosynthesis complex. The structure of AlgX has been shown to have an N-terminal hydrolase domain involved in the acetylation of alginate, and a C-terminal portion where a sugar-binding domain is possibly involved in the binding and orientation of the polymannuronate chain [78, 130]. AlgJ/AlgV and AlgX show about 70% analogy [76, 77], very similar topology [129], and are both required for the esterification of alginate. AlgF has no sequence homology with any other proteins involved in the O-acetylation [10], and *algF*<sup>−</sup> mutants of *A. vinelandii* have been shown to produce non-acetylated alginate [131]. The order of transfer of the acetyl donor to each of the acetylation-related enzymes is uncertain, but the fact is that AlgJ/V performs is directly involved in the O-acetylation of ManA residues and that AlgF serves presumably an accessory role due to the lack of identifiable catalytic residues [129]. Moreover, it is clear that AlgX is also able to catalyze the direct O-acetylation of alginate, receiving the acetate or acetyl donor molecules either from AlgJ/V or AlgF [129]. AlgX has also been shown to interact with MucD in some sort of undetermined regulatory mechanism, which affects the formation of alginate [65, 69, 132].

The esterification of ManA units on hydroxyl groups at C-2 and/or C-3 prevents degradation of the growing alginate polymer by AlgL, an alginate lyase present in the periplasm which will not be discussed in detail herein. O-acetylation also hampers the epimerization of esterified ManA units by AlgG, which is a C5-mannuronan epimerase of the alginate biosynthesis complex and catalyzes the epimerization of β-D-mannuronate (M-blocks) to α-L-guluronate (G-blocks) at polymer level [92]. Non-acetylated ManA units may undergo epimerization by AlgG, which causes the monosaccharide unit to suffer a dramatic change in both its anomeric configuration (from β to α) and pyranose chair conformation (from 4 C1 to 1 C4) (**Figure 4**). Such considerable structural changes have very important effects on the physicochemical properties of the mature alginate depending on the proportion and sequence of M-blocks that go through epimerization [69, 112, 133]. AlgG adopts a right-handed parallel β-helix fold, and the catalytic mechanism proposed for this epimerase is based on the β-elimination (not hydrolysis) reaction of polysaccharide lyases [134], which involves the neutralization of the carboxylate group of the uronic acid, removal of the proton at the C-5 position and cleavage of the glycosidic bond with the formation of a double bond between C-4 and C-5 for the formation of a transient glycal intermediate. For the epimerization reaction, however, glycosidic linkage remains intact, and a proton is added to the opposite side of the C-5 to form the epimer [135]. Functional analysis of AlgG mutants suggests that His319 is the catalytic base and that Arg345 is responsible for neutralizing the carboxylic groups during the epimerase reaction. It has also been proposed that water is the likely catalytic acid [134].

Polymer-level epimerization of monosaccharide units is rarely found in nature, and thus far has only been described for the biosynthesis of alginate (both microbial and algal) and glycosaminoglycans (animal origin) [136]. Hence, AlgG is a very exclusive and extraordinary enzyme found in both *P. aeruginosa* and *A. vinelandii*. These enzymes show an optimum pH for activity between 6.0 and 7.5 and share about 60% sequence identity [134, 135]. However, differently from Pseudomonads, the *A. vinelandii* genome also encodes a family of seven extracellular Ca2+-dependent epimerases (AlgE1–AlgE7), which will be discussed later in this chapter.

### **3.4 Secretion of the mature alginate chain**

The two remaining proteins that are part of the alginate biosynthesis complex are AlgK and AlgE/AlgJ (*Pseudomonas*/*Azotobacter*). AlgK is a lipoprotein of unclear function found anchored into the inner leaflet of the bacterial outer membrane [137]. It is hypothesized that AlgK may have a protective role, guiding the nascent alginate polymer through the biosynthetic multi-protein complex in the periplasmic space, and preventing it from being degraded by lyases, since the lack of AlgK has been shown to result in the secretion of free uronic acids and/or short-chain alginate products [78, 132], showing that AlgK an essential element for the biosynthesis of full-length alginate and the development of the mucoid phenotype in *P. aeruginosa* [138]. Additionally, AlgK is proposed to interact with both Alg44, stabilizing the Alg8-Alg44 polymerase complex, and the porin protein AlgE/AlgJ [112, 137]. There is some evidence to suggest that AlgK is involved in the localization of AlgE/AlgJ to the outer membrane [139]. Also, it has been shown that AlgK may form complexes with AlgX and MucD in the periplasmic space, possibly halting the regulatory role of MucD due to its sequestration by an AlgK/AlgX complex in the biosynthetic multiprotein scaffold [65]. The large number of interactions of AlgK with other proteins may be due to its structure containing several tetratricopeptide-like helical motifs which allow unspecific protein-protein interactions [139].

AlgE was first described in the early 1990s as an integral outer membrane protein strongly associated with the mucoid phenotype of *P. aeruginosa* [140, 141]. Its counterpart in *A. vinelandii*, AlgJ, shares a high degree of similarity and is thought to perform the same function as in Pseudomonads [142]. Although AlgE is classified as a member of the general diffusion porin family [143], biochemical and electrophysiological analyses suggest that this protein is specifically responsible for the passage of alginate into the extracellular environment through the outer membrane, rather than allowing the passage of small molecules and ions, and that it forms a strongly positively charged anion-selective pore that can be partially blocked by GDP-ManA [144, 145]. AlgE/ AlgJ is a monomeric 18-stranded β-barrel protein homologous to the OprD family of substrate-specific porins. In addition to its alginate transport function, it has been shown to play an important role in the assembly of the multiprotein complex involved in alginate biosynthesis [145, 146].

Secretion usually indicates the end of the bacterial alginate biosynthetic pathway, especially in Pseudomonads, and a schematic representation of the entire biosynthetic/regulatory process can be found in **Figure 5**. After release into the environment, alginate may undergo further non-enzymatic structural modifications such as de-O-acetylation and/or depolymerization, albeit to a very modest extent, depending on the chemical conditions in the medium surrounding the cell. However, the alginate produced by *Azotobacter vinelandii* can undergo very significant enzymatic processing extracellularly due to the presence of C5-mannuronan epimerases and alginate lyases, which are also secreted by this bacterium.

### **3.5 Post-secretion chemical modifications of alginate**

### *3.5.1 Extracellular epimerases*

As previously described, *Azotobacter vinelandii* encodes seven different C5-ManA epimerases (AlgE1–AlgE7) that are exported to the extracellular environment and are able to promote polymer-level epimerization of D-ManA units into L-GulA on

**Figure 5.**

*Schematic representation of bacterial alginate biosynthesis, regulation, periplasmic transportation/modification, secretion, and post-secretion modification. Protein structures are merely illustrative (https://pdb101.rcsb.org/) and the nomenclature used is that established for P. aeruginosa encoded genes. (\*A. vinelandii counterparts or enzymes that are exclusively found in the Azotobacter genus are marked with an asterisk).*

alginate. Although the epimerization mechanism for these enzymes is very similar to the mechanism proposed for AlgG, extracellular epimerases have been found to be Ca2+-dependent, in contrast to the periplasmic counterpart. It has been considered that extracellular alginate epimerases are necessary for *A. vinelandii* cells for the creation of a diffusion barrier against O2 [84] as a protective mechanism for its oxygen-sensitive nitrogenase complex and that enzymes of the AlgE family play an important role in the encystment process induced by adverse environmental conditions [147]. Also, only alginate-producing cells may produce cysts that can germinate after storage [65]. Alginates produced by Pseudomonads show lower L-GulA content in comparison to the *A. vinelandii* polymers [38, 65, 70]. Furthermore, alginates isolated from *Azotobacter* spp. contain G-blocks (consecutive L-GulA residues), while in *Pseudomonas* spp. alginates have been found to contain only single L-GulA units on their backbone structure [70]. It is proposed that post-secretion epimerization of alginate by AlgE1–7 is the reason for the higher amounts of L-GulA found in the alginate produced by *A. vinelandii* [10, 65, 118]. As the G-block amounts increase, so does the affinity of this polymer for divalent ions, which are responsible for the formation of cross-links between alginate chains creating a more rigid gel surrounding *A. vinelandii* cells, being physicochemically much different from the lower viscosity alginate produced by bacteria of the genus *Pseudomonas* [10]. Even though consecutive GulA units have not been found in alginates produced by Pseudomonads to date, it has been reported that *P. syringae* pv *glycinea* encodes a C5-mannuronan epimerase designated PsmE, which was found to be able to introduce G-blocks on alginate *in vitro* [65, 148]. Differently from the *A. vinelandii* extracellular epimerases, PsmE is able to epimerize acetylated D-ManA units after removing O-acetyl groups, thus showing a bifunctional deacetylase/epimerase characteristic [148].

### *Bacterial Alginate Biosynthesis and Metabolism DOI: http://dx.doi.org/10.5772/intechopen.109295*

Isomerases of the AlgE family have been shown to be differentially expressed during the life cycle of *A. vinelandii* and to generate different patterns of G-blocks on alginate [147]. Five of the *algE* genes of *A. vinelandii* are found clustered in the genome, while *algE5* and *algE7* are found isolated elsewhere (see **Figure 1**), and each correspondingly encoded enzyme appears to have different specificities in terms of substrate preferences and non-random formation of G- and GM-blocks on the alginate backbone [149]. These extracellular alginate epimerases from *A. vinelandii* are about 70% identical but share less than 10% sequence identity with AlgG [134]. The characterization of AlgEs from *A. vinelandii* has been made at certain levels after cloning, heterogeneous expression, and purification, and three putative AlgE-like isomerases have also been described and characterized in *A. chroococcum* [58]. All of the AlgE isomerases are structurally organized into two types of alternating modules, each with different roles for these enzymes, namely: A-module and R-module (**Table 2**). It is known that the catalytic site is located in the A-modules, each comprising about 385 amino acids (with a primary amino acid sequence homology of ~85%). These are always followed by 1–4 R-modules (~155 amino acids each), which, in turn, are thought to modulate the epimerization rate by stabilizing the process with an extended alginate binding site [150, 151], and have also been shown to play a role in the epimerization pattern of the final alginate product [152]. A-modules have been calculated to bind 11 uronic acid units, while each of the R-modules can bind five units with different binding strengths, which also affects the processivity of a particular AlgE enzyme [151]. Four to seven copies of a nine amino acid long motif are located in the N-terminal region of the R-modules and are known to be Ca2+-binding sites [153]. Calcium ions are important not only for the structural integrity of the protein but also for neutralizing the charges of ManA during the epimerization reaction, in a similar manner Arg345 functions in AlgG [149, 150]. In addition, the last R-module of each extracellular AlgE usually contains an unstructured peptide involved in the secretion of the enzyme [151]. Despite the large sequence homology


*\*Bifunctional extracellular epimerase/lyase enzymes. A-modules: Location of the catalytic site; R-modules: Ca2+-binding sites and substrate-binding/stabilization extended sites.*

#### **Table 2.**

*Structural organization, substrate specificity, and products formed by extracellular C5-ManA epimerases of Azotobacter vinelandii.*

they share, different AlgE epimerases catalyze distinct residue sequences in the alginate product. It has been proposed that the different epimerization patterns depend on the concerted action of both the A- and R-modules, rather than only the catalytic activity of A-modules [150].

Although the extracellular C5-epimerases AlgE1–7 play an important role in the addition of L-GulA units to the alginate produced by *A. vinelandii*, these enzymes have also been cloned and isolated to be used for *in vitro* modification of macroalgal-derived alginate [128, 149]. AlgE1 and AlgE3 are the only epimerases that contain two catalytic A-modules, one of which introduces single G units that form alternating GM sequences, while the other can generate G-blocks with consecutive GulA units [65, 154]. AlgE2 and AlgE5 both consist of an A-module followed by four R-modules, and they primarily introduce short G-blocks on the alginate chain [149]. The same module organization is found in AlgE6, which, however, is able to introduce longer G-blocks on polymannuronic acid [147, 153]. AlgE4 consists of only one A-module and one R-module and is the smallest enzyme among the extracellular C5-epimerases. Although it inserts a high proportion of GulA units into the alginate chain, this enzyme is unable to attach successive G-blocks to the polymer [65, 149]. AlgE7 generally produces the same GM alternating products as AlgE1, AlgE3, and AlgE4. However, AlgE7 and AlgE2 have also been shown to have present lyase activity (breaks G-MM and G-GM bonds), which is thought to share the same active site with epimerase activity [86]. These bifunctional enzymes can degrade the alginate chain after the epimerization reaction, but at different rates. AlgE2 is thought to have low lyase activity and has been shown to perform 1 to 3 elimination reactions per thousand epimerized units [149, 155]. In contrast, AlgE7 has a much more pronounced lyase activity, with a predicted number of 3–4 glycosyl bond breaks per 26 conversions of D-ManA to L-GulA units [86].

Regarding the reaction mechanisms of Ca2+-dependent AlgE epimerases, two reaction modes have been proposed: (1) the preferred attack mode, in which the enzyme preferentially attacks M-units with G residues as the nearest neighbor and detaches from the substrate after each epimerization reaction [58, 149], and (2) the processive mode, in which there is a relative displacement of the enzyme and the alginate chain that allows the conversion of the nearest M-unit without dissociation of the enzyme-substrate complex [149, 153]. AlgE4 has been reported to have a processive mode of action by sliding along the substrate chain and epimerizing every other M residue [156], and this is thought to be due to the 180°-turn between the successive β-(1 → 4)-linked ManA units in the alginate polymer [153]. AlgE2 follows the mechanism of preferred attack mode and is related to the Ca2+ concentration and substrate type of the blocks found in alginate [149, 155]. The remaining AlgE epimerases appear to behave according to either a preferred attack mode or a combination of the two mechanisms [149].
