Molecular mechanisms and genetics of hyaluronan biosynthesis

Michael O'Regan*, Irene Martini, Fabiana Crescenzi, Claudio De Luca and Manfred Lansing Fidia Advanced Biopolymers, via Ponte della Fabbrica, 3/A, 35031 Abano Terme, Italy

Hyaluronan is an extremely important polysaccharide from both the biological and commercial points of view. This review summarizes the present state of the art concerning the polymer and our understanding of the molecular mechanisms of its synthesis with emphasis on the implications of this understanding for polysaccharide engineering of hyaluronan.

Keywords: hyaluronan; biosynthesis; polysaccharideengineering

Hyaluronan (HA, Figure 1) is a naturally occurring linear polysaccharide composed of repeating disaccharide units of glucuronic acid and N-acetyl glucosamine linked by fl-l-3 and fl-l-4 glycosidic bonds. HA is present in all soft tissues of higher organisms, and in particularly high concentrations in the synovial fluid and vitreous humour of the eye 1. In addition to fulfilling structural roles related to its lubricating and water-retaining properties, evidence is mounting that HA also plays an important role in a number of biological processes such as cell motility and cell-cell interactions (see refs 2 and 3 for recent reviews). Since the discovery of HA, much attention has focussed on the possible biomedical applications of highly purified HA fractions. Numerous studies have proven that HA isolated from various sources has an identical chemical structure. Therefore, since it is already present in the human body, exogenously applied HA of sufficient purity is highly biocompatible (i.e. it does not provoke adverse host reactions) and completely biodegradable by natural catabolic pathways. Solutions of purified HA of high molecular weight are extremely viscous and demonstrate very interesting rheological behaviour. As a consequence of the above properties, HA has been widely exploited in the fields of viscosupplementation and viscosurgery 2. Widespread exploitation of HA in other fields where its natural properties might have rendered it suitable has been limited by the fact that unmodified HA exists only in the form of an aqueous gel and has a short residence time upon administration. Therefore, much attention has been focussed on obtaining chemical derivatives of HA which would maintain the biocompatibility of the parent molecule while allowing it to be processed into products for use in areas such as drug delivery and tissue repair 4'5.
* To whom correspondence should be addressed

Sources of hyaluronan
In order to fulfil the raw material requirements for products for the biomedical applications described above, the identification of dependable and economically viable sources of HA has been an industrial priority. Initially, attention was focussed on the extraction of HA from animal tissues. Numerous tissue sources, including umbilical cord, skin and rooster combs, have been evaluated. Subsequently, rooster comb HA became the most widely used and traditionally accepted source both from an industrial and regulatory point of view. However, there are certain drawbacks to dependence on this source of HA. High-molecular-weight material is difficult and costly to isolate due to the fact that the HA is complexed with proteoglycans. Additionally, animalsourced materials for biomedical applications are coming under increasingly stringent regulatory control due to the fear of contamination with both conventional and unconventional viral agents. Finally, if HA-based products become commonly used in sectors such as drug delivery and tissue repair, predictions indicate that rooster comb supplies will be insufficient to meet the demand for HA. Therefore, attention has turned in recent years to the identification of alternative HA sources. Lancefield's group A and C streptococci, which are human and animal pathogens respectively, produce HA and have been exploited for the development of industrial-scale fermentation processes (for example, US patents 4784990 and 4517295). The equivalence of streptococcal and rooster comb HA has been demonstrated, and the former has now been accepted from a regulatory point of view. Supplies of HA from streptococcal fermentation are theoretically limitless, and no fears of seasonal fluctuations or batch-to-batch variations exist if a tightly controlled process is used. The

0141-8130/94/060283-04

~? 1994Butterworth-HeinemannLimited

Int. J. Biol. Macromol. Volume 16 Number 6 1994 283

Molecular mechanismsof hyaluronan biosynthesis."M O'Regan et al.

o H
C-O-H CH 2 OH 0 C-O -H CH OH 2

._".1

.J

oo.

OH

OH

" ? " r
||

"
0

?""I"

"
n-1

NHCCH

II
o

Figure 1 Structure of hyaluronan

HA is produced as a capsule of high-molecularweight material which is easily separated from the micro-organisms, although extensive purification is still necessary in order to obtain material of sufficient purity. The possibility of contamination with viral agents is obviously not a major concern, and the use of nonpathogenic mutants in the industrial processes ensures the absence of other toxic impurities. An added advantage of using streptococci, which has become apparent in recent years, is that recombinant DNA technology can be applied to these organisms. Therefore, the tools of molecular biology can be exploited to facilitate the development of a clear understanding of the molecular mechanisms of HA synthesis and to allow intervention by genetic engineering to modulate this process.

GLUCOSE

~ hexokinaseglucose -6-phosphate isomerase

GLUCOSE-6-P ~ FRUCTOSE-6-P

phospho glucomutase

glucosamine-6-P synthase
GLUCOSAMINE-6-P

GLUCOSE-1-P

mutase GLUCOSAMINE-I-P glucosamine-l-P acetyl transferase

N-ACETYL GLUCOSAMINE-1-P

UDP-glucose 1-P-uridyltransferase (pyrophosphorylase)

UDP-GLUCOSE

Biosynthetic pathway
An important step in opening up the possibilities of genetic engineering is a definition of the HA biosynthetic pathway. Although no detailed genetic work has been carried out to identify each step in the biosynthetic pathway in streptococci, the succession of biochemical events can be put together based on knowledge of the biosynthesis of the two UDP-sugar precursors of HA gained from studies carried out in other organisms 6'7 and from some limited streptococcal studies 8. The proposed biosynthetic pathway for HA is shown in Figure 2. The precursors, UDP-GIcA and UDP-GIcNAc, are synthesized as side reactions of the glycolytic pathway starting from glucose-6-phosphate and fructose6-phosphate, respectively. Experiments carried out with radioactive glucose (O'Regan, unpublished data) have shown that about 5-7% of the glucose in the streptococcal culture medium is converted to HA (Table 1). Metabolic engineering approaches designed to increase the flux towards HA could be fruitful in terms of obtaining yields that are more industrially viable.

UDP-glucose dehydrogenase

N-acetyl glucosamine 1-P-uridyl transferase (pyrophosphorylase)

UDP-N-ACETYL GLUCOSAMINE

UDP-GLUCURONIC ACID

hyaluronan synthase

HYALURONAN
Figure 2 The proposed biosynthetic pathway for hyaluronan

Table I

Distribution of t4C label in various culture fractions following growth of a group C HA-producing streptococcus in the presence of 14C-labelled glucose Radioactivity (% of supplied 14C) Fraction Spent medium + wash solutions Cell pellet HA 10 h culture 93.5 1 5.5 14 h culture 88.8 4.2 7

Hyaluronan synthase
Much attention has been focussed on hyaluronan synthase as a key enzyme in the biosynthetic pathway of HA and an essential element in developing an understanding of the mechanism of synthesis. HA synthase is located in the plasma membrane. This has been shown by a number of different workers using various approaches. Markovitz and Dorfman 9 first synthesized HA using streptococcal membranes. Prehm 1,~ 1 demonstrated the extracellular growth of the HA chain and that HA was not synthesized in the Golgi

apparatus as are other members of the glycosaminoglycan family. Subsequently, Prehm ~2 showed that, in the presence of added U D P precursors, HA synthesis was ten times greater in disrupted F9 cells as compared to intact cells due to the increased accessability of the

284

Int. J. Biol. Macromol. Volume 1 6 Number 6 1994

Molecular mechanisms of hyaluronan biosynthesis: M. O'Regan et al. precursor pool. This suggests that the active domain of the HA synthase may be located on the internal surface of the plasma membrane as has been recently hypothesized for the bacterial enzyme 13 Many attempts have been made to identify the peptide components of both the prokaryotic and eukaryotic HA synthases. In streptococci, Prehm 14 has found proteins of 75, 52, 47, 42 and 34 kDa in active HA synthase preparations. Photoaffinity labelling is revealed three proteins of 42, 33 and 27kDa which could bind UDP-GlcA. In eukaryotes, a number of different groups have identified active complexes containing several peptide components ranging in size from 116 to 52 kDal 6-19. It can be concluded from the work described above that a complex of plasma-membrane-bound proteins is involved in the synthesis and transport of HA in both prokaryotes and eukaryotes. However, none of the above approaches allowed clear identification of the HA synthase. Recent work in streptcocci, involving a transposon mutagenesis approach, has resulted in identification of the prokaryotic HA synthase as a 42 kDa membrane protein (see below for details). This significant advance should permit an improvement in the understanding of the molecular mechanisms of HA synthesis in streptococci and a subsequent transposition of this knowledge to eukaryotic organisms. A number of unusual features of HA synthesis have been revealed in the course of studies by various laboratories. If one protein is responsible for HA synthesis, then the synthase must contain independent binding sites for each substrate. Polymerization is rapid, having been estimated using membrane fractions at approximately 100 sugar units per minute in eukaryotesl. HA synthesis appears to occur by alternate transfer of the growing chain to UDP-GIcA and UDP-GIcNAc. Unusually, chain growth seems to occur at the reducing end t. Despite numerous attempts, no evidence has been obtained for the involvement of a lipid intermediate in HA synthesis2, as is also the case for cellulose synthesis in Acetobacter xylinum 21. Identification of genetic components As pointed out above, efforts towards identification of the genes involved in HA synthesis have concentrated on streptococci due to the greater amenability of these organisms to genetic manipulation. Two approaches have been taken to identify the HA biosynthetic genes. Prehm and co-workers have attempted to optimize the purification of group C streptococcal membrane fractions possessing HA synthase activity and have identified the protein components. One of these was a 52kDa protein which was shown by a number of experimental approaches to be implicated in HA synthesis 14. The production of antibodies against this protein permitted its cloning from a streptococcal gene bank 22. The protein showed amino acid sequence similarity to a number of bacterial proteins involved in oligopeptide binding and transport, suggesting that the protein may be involved in transport of HA. Further evidence for this hypothesis has been obtained from experiments in which the gene was over-expressed on a plasmid in streptococci (see below). A number of other workers have approached the problem by isolating stable non-mucoid mutants of streptococci by transposon mutagenesis and subsequently isolating and characterizing the DNA flanking the inserted transposon 2a-25. This work led to the identification, in group A streptococci, of an operon containing three genes involved in HA synthesis: hasA, B and C. The protein product of hasA is a membrane protein and shows amino acid sequence similarity to chitin synthase and the nodC product of rhizobium which is involved in oligosaccharide synthesis la'26. These results, combined with the finding that membranes from Escherichia coli clones expressing hasA possess HA synthase activity, suggest that hasA encodes the HA synthase. Studies on the remaining two genes of the operon have shown that hasB and hasC encode UDP-GIc dehydrogenase and UDP-GIc pyrophosphorylase, respectively26'27. The present state of the art concerning the genes involved in HA synthesis is summarized in Figure 3. The has operon encodes genes involved in the biosynthesis of UDP-GIcA and in the polymerization of HA. The second precursor, UDP-GIcNAc, is probably synthesized by so-called 'housekeeping' genes, as this sugar derivative is also essential for cell-wall synthesis. Evidence to date indicates that the 52 kDa protein is implicated in HA transport. Advances need to be made in determining which factors regulate expression of these genes and determine the division of the UDP-GIcNAc pool between cell-wall and HA synthesis. An important objective is also to determine the similarity of the genes from eukaryotic systems and those identified to date in bacteria. Prehm and co-workers have identified a protein of approximately 52 kDa from fibroblasts which crossreacts with antibodies against the streptococcal 52 kDa polypeptide 19. No function has yet been assigned to this protein in the eukaryotic systems.

Recombinant HA-producing strains

A number of attempts have been made to construct recombinant HA-producing strains and assess the advantages of these strains with respect to the parent organisms. In our laboratory, temperature-sensitive expression vectors have been developed for HA-producing streptococci, and the gene encoding the 52 kDa protein has been expressed in these organisms (Martini et al., manuscript in preparation). The results are summarized in Table 2 and show that strains expressing this gene on a plasmid produce higher levels of HA and that this HA has a lower molecular weight than that produced by the wild-type p52 "has ABC v transport?

synthesis G l c N A c
"~

precursor biosynthesis and polymerization HA

Figure 3 Genes involved in the biosynthesis of hyaluronan. The gene encoding the 52 kDa protein was cloned from a group C streptococcal gene bank using antibodies raised against the purified protein. The hasABC operon was identified by characterizing the DNA segments flanking transposon insertion sites in HA-negative mutants of group A streptococci. The genes for UDP-GlcNAc biosynthesis have not yet been identified

Int. J. Biol. Macromol. Volume 16 Number 6 1994

285

Molecular mechanisms of hyaluronan biosynthesis: M. O'Regan et al. Table2 The effectof elevatedcopy number of the gene encoding the
52 kDa protein on the quantity and molecular weight of HA produced by Streptococcus equi
S. equi S. equi

which could either be a highly purified streptococcal membrane fraction or a purified HA synthase obtained from a recombinant heterologous host.

Parameter HA (mg ml- 1) mg HA/rag cells Molecular mass (kDa)

pY09a 1299 49.9 2000

pY016= 1591 66.3 4-500

Conclusions
Considerable progress has been made in recent years in our understanding of the molecular mechanisms of HA synthesis. Although the system is not yet completely understood, the identification of a number of gene components involved in HA synthesis has certainlygiven added impetus to the possibility of unravelling the remaining mysteries surrounding the synthesis of HA. A number of workers have constructed recombinant HA-producing strains and results have shown that polymers with modified characteristics can be obtained. However, attention should be paid to the regulatory status of HA obtained from recombinant strains. Even though this molecule has the same chemical structure as HA from non-genetically modified organisms, certain regulatory authorities would consider it a new molecule (thus obliging a re-evaluation of its toxicology) simply because a recombinant strain is being employed. Preliminary studies indicate that the industrial production of HA by in vitro enzymatic synthesis is a feasible objective. This approach will certainly yield advantages with respect to product characteristics and also in the polysaccharide engineering of HA.

apY09 is a temperature-sensitiveexpressionvector which replicatesin streptococci;pY016is pY09 containingthe genefor the 52 kDa protein under the control of the temperature-sensitivepromoter (Martini et al., manuscript in preparation) strain containing the plasmid vector alone. These results constitute additional evidence that the 52 kDa protein is involved in HA transport. The results also confirm the utility of a genetic engineering approach to polysaccharide engineering which allows the physico-chemical characteristics of the product to be modulated. Other workers have transferred either all or parts of the HA synthase operon to a number of heterologous hosts and shown that HA synthesis at significant levels can be achieved in these hosts 13'25. Further work in this direction might open up the possibility of obtaining polymers with altered characteristics, or, indeed, at lower cost due to the less stringent nutritional requirements of the heterologous hosts.

In dtro enzymatic synthesis of HA

Recent advances in understanding of the HA biosynthetic apparatus have opened up the possibility of developing a system for in vitro enzymatic synthesis of HA. As discussed above, reliable sources of HA exist and are extensively exploited in industry. However, despite the extensive purification carried out on the polymer, concerns are still being expressed about the possibility of contamination with unknown agents such as nonconventional viruses. Additionally, as a consequence of this extensive purification, a polymer of considerable molecular weight polydispersity is obtained. In vitro enzymatic synthesis would permit a polymer of extremely high purity and optimized physico-chemical characteristics to be obtained. The latter properties could be optimized by synthesizing a polymer of the desired molecular weight with a minimized molecular weight polydispersity. Optimization of the technology could also permit synthesis of monodisperse HA oligosaccharides which might demonstrate improved biological activity when compared to the oligosaccharide fractions obtained by hyaluronidase digestion of HA followed by chromatographic separation. In the longer term, the optimization of in vitro technology could permit synthesis of novel polymers by modifying the catalytic site of the synthase to obtain enzymes which combine the sugar moieties in varying ways or incorporate alternative sugar components. Recent work in our laboratory has focussed on the optimization of a process for in v i t r o synthesis of HA. Preliminary studies have shown that the molecular weight and molecular weight polydispersity can be controlled by varying the reaction conditions (Lansing et al., manuscript in preparation). Present work is aimed at identification of the HA synthase with the highest activity,

References
1 2 3 4 Laurent,U B.G.and Reed,R.K.Adv. Drug Deliv. gev. 1991,7, 237 Laurent,T.C. and Fraser, J.R.E. FASEB J. 1993,6, 2397 Knudson,C.B. and Knudson, W. FASEB J. 1993,7, 1233 Rastrelli, A., Beccaro, M., Biviano, F., Calderini, G. and Pastorello, A. in 'Clinical Implant Materials' (Eds G. Heimke, U. Soltrs and A.J.C. Lee)ElsevierSciencePublishers, Amsterdam, 1990, p 199 Balazs, E.A., Bland, P.A., Denlinger, J.L., Goldman, A.I., Larsen, N.E., Leshchiner,E.A. and Morales, B. Blood Coagulation Fibrinolysis 1991, 2, 173 Brede,G., Fjaervik, E. and Valla, S. J. Bacteriol. 1991, 173 (21), 7042 Mengin-Lecreulx,D. and Van Heijenoort, J. J. Bacteriol. 1993, 175 (19), 6150 Matsubara, C., Kajiwara, M., Akasaka, H. and Haze, S. Chem. Pharm. Bull. 1991, 39 (9), 2446 Markovitz,A. and Dorfman,A. J, Biol. Chem. 1962,238(2),273 Prehm,P. Biochem. J. 1983,211, 181 Prehm,P. Biochem. J. 1983,211, 191 Prehm,P. Biochem. J. 1984,220, 597 DeAngelis,P.L., Papaconstantinou, J. and Weigel, P.H.J. Biol, Chem. 1993,268 (26), 19181 Prehm,P. and Mausolf, A. Biochem. J. 1986,235, 887 Van de Rijn,I. and Drake, R.R.J. BioL Chem. 1992,267(34),24302 Mian,N. Biochem. J. 1986,237, 333 Mian,N. Biochem. J. 1986,237, 343

5 6 7 8 9 10 11 12 13

14 15 16 17 18 Ng, K.F.andSchwartz, N.B.J. Biol. Chem. 1989,264(20),l1776 19 Klewes,L.,Turley,E.A.and Prehm,P. Biochem. J. 1993,290,791 20 Sugahara, K., Schwartz, N.B. and Dorfman, A. J. Biol. Chem. 1979, 254 (14), 6252 21 Whitfield,C. Can. J. Microbiol. 1988,34, 415 22 Lansing, M., Lellig, S., Mausolf, A., Martini, 1., Crescenzi, F., O'Regan, M. and Prehm, P. Biochem. J. 1993,289, 179 23 Wessels, M.R., Moses, A.E., Goldberg, J.B. and DiCesare, T.J. Proe. Natl Acad. Sci. USA 1991,88, 8317 24 Dougherty, B.A.andvandeRijn, l.J. Exp. Med. 1992,175,1291 25 DeAngelis,P.L., Papaconstantinou, J. and Weigel, P.H.J. Biol. Chem. 1993,268 (20), 14568 26 Dougherty, B.A.andvandeRijn, l.J. Biol. Chem. 1994,269(1),169 27 Dougherty,B.A.and van de Rijn, I. J. Biol. Chem. 1993,268 (10), 7118

286

Int. J. Biol. Macromol. Volume 16 Number 6 1994