BioEngineering

Metabolic engineering of hyaluronic acid production

Natural sources of hyaluronic acid and its uses

Hyaluronic acid (HA) is a naturally occurring biopolymer, which serves important biological functions in bacteria and higher animals including humans. Naturally occurring HA may be found in the tissue of higher animals, in particular as intercellular space filler (Balazs, 1993). It is found in greatest concentrations in the vitreous humour of the eye and in the synovial fluid of articular joints (O'Regan et al., 1994). In gram positive streptococci it appears as a mucoid capsule surrounding the bacterium (Fig. 1).

Figure 1 Cross-sectioned Streptococcus zooepidemicus cell with their HA capsule from aerated culture (Goh, 1998).

Since its discovery in human tissue, HA and its derivatives has been largely studied and applied in the biomedical arena. The appeal of this polymer has been accentuated by its high level of biocompatility. It has been used in viscosurgery to allow surgeons to safely create space between tissues. As a microcapsule it can be used for targeted drug delivery. Viscosurgical implants are constructed from HA (Balazs, 1993). Its viscoelastic character has been used to supplement the lubrication in arthritic joints. Finally, because of its high water retention capacity, this EPS (extracellular polysaccharide) also occupies a niche in the lucrative cosmetics market.

The commercial value of HA far exceeds that of other microbial EPS. With an estimated world market value of $US 500 million, it is sold for up to $US 100 000 per kilogram. Compare this with another leading microbial EPS, xanthan gum derived from Xanthomonas campestris, which sells for up to $US 11 per kilogram.

Structural features and properties of hyaluronic acid

The utility of this biopolymer is derived from a remarkably simple construct. HA is comprised of linear, unbranching, polyanionic disaccharide units consisting of glucuronic acid (GlcUA) an N-acetyl glucosamine (GlcNAc) joined alternately by beta 1-3 and beta 1-4 glycosidic bonds (Fig. 2). It is a member of the glycosaminoglycan family which includes chondroitin sulphate, dermatin sulphate and heparan sulphate. Unlike other members of this family, it is not found covalently bound to proteins.

Figure 2 Disaccharide repeating unit of HA comprising GlcUA and GlcNAc.

When incorporated into a neutral aqueous solution hydrogen bond formation occurs between water molecules and adjacent carboxyl and N-acetyl groups. This imparts a conformational stiffness to the polymer, which limits its flexibility. The hydrogen bond formation results in the unique water-binding and retention capacity of the polymer. It also follows that the water-binding capacity is directly related to the molecular weight of the molecule. Up to six liters of water may be bound per gram of HA (Sutherland, 1998).

HA solutions are characteristically viscoelastic and pseudoplastic. This rheology is found even in very dilute solutions of the polymer where very viscous gels are formed. The viscoelastic property of HA solutions which is important in its use as a biomaterial is controlled by the concentration and molecular weight of the HA chains. The molecular weight of HA from different sources is polydisperse and highly variable ranging from 10⁴ to 10⁷ Da. The extrusion of HA through the cell membrane as it is produced permits unconstrained polymer elongation and hence a very high molecular weight molecule.

Hyaluronic acid production by bacterial fermentation

HA has been traditionally extracted from rooster combs and bovine vitreous humor. However it is difficult to isolate high molecular weight HA economically from these sources because it forms a complex with proteoglycans (O'Regan et al., 1994). It is presently impractical to control the molecular weight of the biopolymer while it is synthesized in animal tissue. Subsequent extraction and purification processes result in an inherent molecular weight reduction. From a social viewpoint, the use of animal-derived biochemicals for human therapeutics is being met with growing resistance, besides ethic arguments, because of the risk of viral infection. Industry has instead turned to bacterial fermentation processes with the hope of obtaining commercially viable biopolymer. Here the EPS is released into the growth medium and control of polymer characteristics and product yields are feasible. The amount of biopolymer that can be produced by this route is theoretically unlimited.

The use of bacterial fermentation does however come with its own disadvantages. The recent trend has been to use Lancefield’s group A and C streptococci which naturally produce a mucoid capsule of HA. The HA capsule is a biocompatibility factor which enables this gram positive bacteria to evade host immune defenses and hence accounts for its characteristically high virulence level. It is therefore not unreasonable to question the use of such pathogenic bacteria as production strains.

Streptococcal fermentations have only been able to produce HA with an average molecular weight in the range of 1 to 4 MDa. As previously highlighted, high molecular weight enhances the desirable properties of the biopolymer. Given that 10 MDa chains can be obtained from animal tissue, there is considerable scope for improvement.

Streptococci are nutritionally fastidious, facultative anaerobes which produce lactic acid as a by-product of glucose catabolism. Hence the energy recovered by these bacteria is lower relative to aerobic bacteria. The yield of HA from bacterial fermentation to date has been characteristically low (0.1 g/g glucose, 0.15 g/lh) and would certainly struggle to meet market demand. The strict nutritional requirements influences the fermentation economics by prohibiting the use of chemically defined media for production scale fermentations and limits the choice of complex media that can be employed.

Hyaluronic acid fermentation research at the University of Queensland

Research in the Bioprocess group at the Department of Chemical Engineering has studied HA production in Streptococcus zooepidemicus since the early 1990’s. Initially our research focused upon process parameter optimization, selection of process mode, HA yield optimization by complex medium design, and model based fermentation control (Lertwerawat, 1993; Johns et al., 1994; Armstrong et al., 1997; Goh, 1998). As the importance of HA molecular weight became increasingly clear, the group has recently turned to the effect of process parameters on the molecular weight properties of HA (Armstrong, 1997; Armstrong and Johns, 1997; Armstrong and Johns, 1995) as a major area of study.

Several fermentation parameters were found to significantly affect the molecular weight of HA produced while others had little influence. Specifically, low growth temperatures (28^oC), culture aeration and high initial glucose concentration (40 g/l) resulted in the production of higher molecular weight HA in S. zooepidemicus. Culture pH and agitation did not influence the molecular weight outcome of HA fermentations.

Figure 3 HA biosynthetic pathway in streptococci.

Common to these observations has been an inverse relationship between the specific growth rate of the bacteria and the molecular weight of HA produced. This effect has been explained by a resource-based metabolic model for HA synthesis. This model is illustrated in Figure 3. In its most simplistic form we can identify two competing processes within the bacterial cell. These two processes are cell growth and the biosynthesis of HA. These two processes compete for the limited resources namely carbon, nitrogen and energy. At low specific growth rates the cell directs more glucose-derived activated precursors (namely UDP-Glc and UDP-GlcNAc) to HA synthesis rather than cell wall synthesis. The higher ATP yields from aerobic glucose catabolism favors the formation of UTP, which is required for the formation of the two activated precursors of HA, synthesis (UDP-GlcUA and UDP-GlcNAc). Glucose, which may be used to synthesise HA, is also depleted by lactate production under anaerobic growth.

The specific rate of HA production (g g^-1 h^-1) was also observed to increase with decreasing specific growth rate. Given the density of active HA-synthesising enzymes is unlikely to change, the higher rate of production can be attributed to a higher polymerisation rate through each synthase. An increased synthase activity may arise from the higher intracellular substrate concentration resulting from low growth rate.

Kitchen and Cysyk (1995) demonstrated that HA synthase derived from fibroblast cells have a limited half-life of 2 to 4 hours. The similarity between the time it takes to produce one HA chain and the lifespan of the synthase led them to propose the enzyme synthesises only one HA chain during its lifetime. This tells us that the molecular weight of the HA produced is determined by the number of precursors that the synthase is able to polymerise during its lifetime. Based upon this theory, overexpression of the synthase will not increase the molecular weight of HA synthesised. Indeed it may reduce the mean molecular weight since there are more enzymes competing for the same resources. This effect was reported for the overexpression of polyhydroxybutyrate synthase in recombinant E. coli (Sim et al., 1997).

The foregoing discussion leads us to speculate upon several strategies for increasing the molecular weight of HA produced.

• increase the lifespan of the synthase
• decrease the rate of biomass synthesis to reduce resource competition
• increase the resource flux towards HA synthesis
• increase the energy efficiency of the cell

Figure 4 Organization of the has operon from S. pyogenes. has A coding the HA synthase; has B coding the UDP-Glucose dehydrogenase and has C coding the Glucose-1-P-Pyrophosphorylase. The size of the operon is 3606 bp.

Heterologous gene expression studies in E. coli and S. cerevisiae showed that only HA synthase is necessary for HA production. HA synthases from several species including human, mouse, frog, virus and other bacteria besides streptococci are known. The synthases are highly conserved across many species (Fig. 5) with the notable exception of the genes from Pasteurella multocida and chlorella virus PBCV-1. Although the synthases from many species are homologous, the polymerization rate differs greatly between the synthases. The HA synthase from S. equi has the highest yet reported polymerisation rate (160 monosaccharides per second) (Kumari and Weigel, 1997).

Figure 5 Evolutionary relationship among hyaluronan synthases. Group C streptococci seHAS, group A streptococci spHAS, murine muHAS1-3, human huHAS1-2 and frog xlHAS. (Kumari and Weigel, 1997).

Enzymes from the metabolic pathway of UDP-N-Acetylglucosamine, the second precursor of HA synthesis have not been characterized in streptococci. However UDP-N-Acetylglucosamine is also a precursor of peptidoglycan (Fig. 3) and wall polysaccharides therefore the existence of enzymes in this pathway is known from studies in other bacteria.

Future research initiatives

Future research by this group will look further into the resource-based model for HA molecular weight control. The tools of metabolic engineering will be employed to develop a rational strategy for improving HA yields and molecular weight from bacterial-based fermentation. Specifically the following areas will be targeted.

• Development of a metabolic flux model for HA biosynthesis in S. zooepidemicus that will enable us to define the boundaries of the resource allocation model.
• Cloning the has operon from S. zooepidemicus to allow us to engineer a HA-producing bacterial strain that is not reliant upon fermentative catabolism.
• Simply overexpressing the has operon is unlikely to produce high molecular weights. Metabolic control analysis will be employed to assess the merits of varying the relative expression level of the three genes in the has operon. Random mutagenesis of the promoter region will provide data for performing the MCA.

• American Type Culture Collection
• Australian National Genomic Information Service
• DTU Copenhagen, Center for Process Biotechnology
• Journal of Biological Chemistry