Biotechnology Advances 25 (2007) 537 – 557
`
`www.elsevier.com/locate/biotechadv
`
`Research review paper
`The many ways to cleave hyaluronan
`Robert Stern a, Grigorij Kogan b,⁎, Mark J. Jedrzejas c, Ladislav Šoltés d
`
`a Department of Pathology, School of Medicine and UCSF Comprehensive Cancer Center,
`University of California San Francisco, San Francisco, California 94143-0511, USA
`b Institute of Chemistry, Center for Glycomics, Slovak Academy of Sciences, SK-84538 Bratislava, Slovakia
`c Center for Immunobiology and Vaccine Development, Children's Hospital Oakland Research Institute, Oakland, California 94609, USA
`d Institute of Experimental Pharmacology, Slovak Academy of Sciences, SK-84104 Bratislava, Slovakia
`
`Received 18 May 2007; received in revised form 6 July 2007; accepted 11 July 2007
`Available online 18 July 2007
`
`Abstract
`
`Hyaluronan is being used increasingly as a component of artificial matrices and in bioengineering for tissue scaffolding. The length
`of hyaluronan polymer chains is now recognized as informational, involving a wide variety of size-specific functions. Inadvertent
`scission of hyaluronan can occur during the process of preparation. On the other hand, certain size-specific hyaluronan fragments may
`be desirable, endowing the finished bioengineered product with specific properties. In this review, the vast arrays of reactions that
`cause scission of hyaluronan polymers is presented, including those on an enzymatic, free radical, and chemical basis.
`© 2007 Elsevier Inc. All rights reserved.
`
`Keywords: Hyaluronan; Hyaluronidase; Reactive oxygen species; Degradation; Fragmentation; Hydrolysis
`
`Contents
`
`1.
`2.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Introduction .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Enzymatic catabolism of hyaluronan .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.1. Historical background .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.2.
`Prokaryotic enzymes that cleave HA .
`Bacterial β-endoglycosidases and their lyase mechanism of action .
`2.2.1.
`Bacterial β-exoglycosidases .
`2.2.2.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Eukaryotic enzymes that cleave HA .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`β-Endoglycosidases .
`2.3.1.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`β-Exoglycosidases .
`2.3.2.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Substrate specificity of the hyaluronidases .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Provisos for the hyaluronidase reactions .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`The transglycosylation reaction .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Inhibitors of hyaluronidases .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`2.4.
`2.5.
`2.6.
`2.7.
`
`2.3.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`. 538
`. 539
`. 539
`. 539
`. 539
`. 540
`. 540
`. 540
`. 541
`. 542
`. 542
`. 542
`. 542
`
`⁎ Corresponding author. Present address: Directorate General Research, European Commission, B-1050, Brussels, Belgium. Tel.: +32 2 2983113;
`fax: +32 2 2955365.
`E-mail address: grigorij.kogan@ec.europa.eu (G. Kogan).
`
`0734-9750/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
`doi:10.1016/j.biotechadv.2007.07.001
`
`Exhibit 1056
`Prollenium v. Allergan
`
`

`

`538
`
`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`.
`.
`.
`3. Non-enzymatic reactions that degrade HA .
`.
`.
`.
`3.1. Acidic and alkaline hydrolysis .
`.
`.
`.
`.
`.
`.
`3.2. Ultrasonic degradation .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.
`Thermal degradation .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.4. Degradation by oxidants .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.4.1.
`Superoxide anion radical .
`.
`.
`.
`.
`3.4.2. Hydrogen peroxide .
`.
`.
`.
`.
`.
`.
`.
`3.4.3.
`Singlet oxygen .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.4.4. Hydroxyl radical .
`.
`.
`.
`.
`.
`.
`.
`.
`3.4.5. Nitric oxide .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.4.6.
`Peroxynitrite anion .
`.
`.
`.
`.
`.
`.
`.
`3.4.7. Hypochlorous acid .
`.
`.
`.
`.
`.
`.
`.
`3.4.8.
`Carbonate radical anion.
`.
`.
`.
`.
`.
`3.4.9. Dichloride radical anion .
`.
`.
`.
`.
`3.5. Miscellaneous degradations .
`.
`.
`.
`.
`Preparation and application of hyaluronan fragments .
`4.
`5. Concluding remarks .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Acknowledgements .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`References .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`543
`543
`543
`544
`544
`546
`546
`547
`547
`547
`548
`549
`550
`551
`551
`551
`553
`553
`553
`
`1. Introduction
`
`Hyaluronan (HA) is a high-molar-mass linear glycos-
`aminoglycan (GAG) found in the extracellular matrix
`(ECM). This linear polysaccharide can reach a size of 6 to
`8 MDa. It is ubiquitous, but is particularly prominent in
`tissues undergoing rapid growth and repair. The polymer
`has the structure of poly[(1 → 3)-2-acetamido-2-deoxy-
`β-D-glucose-(1 → 4)-β-D-glucopyranosyluronic acid]
`(Fig. 1). It has one carboxyl group per disaccharide
`repeating unit, and is therefore a polyelectrolyte with a
`negative charge at neutral pH. It is near perfect in
`chemical repeats, with no known deviations in its
`simple disaccharide structure with the possible excep-
`tion of occasional deacetylated glucosamine residues.
`High-molar-mass forms of HA are reflections of intact
`normal tissue, while fragmented forms of HA are usually
`indications of stress. Indeed, various size HA fractions
`have an enormous repertoire of functions and constitute
`an information-rich system (Stern et al., 2006). Because of
`its novel characteristics of biocompatibility and rheolog-
`ical properties, HA is being used increasingly as a
`component of artificial matrices and in bioengineering for
`
`tissue scaffolding. The size of HA polymers used in
`bioengineering and biotechnology is obviously of critical
`importance. Inadvertent degradation of the polymer
`during the process of its formulation, cross-linking, and
`other covalent modifications prior to incorporation into
`scaffolding and matrix devices may have deleterious
`effects and limit usefulness of the product.
`Increased evidence has been gathered that low-molar-
`mass HA fragments have different activities than the
`native polymer. Large matrix polymers of HA are space-
`filling, anti-angiogenic, and immunosuppressive, whereas
`the intermediate-sized polymers comprising 25–50 dis-
`accharides are inflammatory,
`immunostimulatory, and
`highly angiogenic. These low-molar-mass oligosacchar-
`ides (OSs) appear to function as endogenous danger
`signals (Kogan et al., 2007, in press). Certain activities of
`HA oligomers are associated with a specific size of the
`fragment and some functions are very sensitive even to
`subtle changes in the chemical structure of the HA
`fragment employed. For example, it has been demonstrat-
`ed that a mixture of HA tetrasaccharide and hexasacchar-
`ide can induce complete maturation of human dendritic
`cells through a Toll-like receptor 4-mediated pathway.
`
`Fig. 1. Fragment of the HA structure.
`
`

`

`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`539
`
`This is in contrast to high-molar-mass HA that does not
`possess such an effect (Termeer et al., 2000).
`In a canine arthritis model, stressed synovial cells
`suppress cell death upon treatment with HA tetrasacchar-
`ides but not with disaccharides or hexasaccharides (Xu
`et al., 2002). Fine variations in the chemical structure can
`also have pronounced effect on the biological properties
`of HA fragments: HA oligomers produced by digestion of
`the polymer with vertebrate hyaluronidase elicit signifi-
`cantly greater release of interleukin-12 from leukocytes
`than do OSs produced by the action of bacterial
`hyaluronan lyases, though the differences in molecular
`structure are rather minor (Jobe et al., 2003). Thus,
`method of preparation of the HA oligomers may affect
`their size and physico-chemical properties as well as
`influence their biological activities and properties.
`For this reason, we have assembled here the array of
`known mechanisms utilized for the cleavage of the HA
`polymer, including both enzymatic and non-enzymatic
`reactions. We attempt to define the limits for working with
`HA, pointing out limitations, defining situations and
`reagents to avoid in order to prevent HA chain scission,
`and to formulate the precise environment within which it
`is possible to work.
`The enzymes are also tabulated so that they can be used
`as tools to trim and fashion polymers, to generate precise
`sizes, as well as to provide desired reducing and non-
`reducing termini. This may appear to be inordinately
`refined. However, it is also becoming apparent that such
`resolution may be necessary. HA fragments of an
`intermediate size can induce apoptosis in cultured cells
`(Alaniz et al., 2006), and inflammation in vivo (Noble,
`2002), while smaller fragments, in the tetrasaccharide
`range ameliorate such effects by suppressing apoptosis,
`and by inducing heat shock proteins (Xu et al., 2002). It is
`entirely reasonable that such small fragments will have
`effects that are dependent on the nature of the residues that
`constitute the termini. Such details were not deemed
`important heretofore, but with the myriad of functions
`ascribed to HA fragments, attention to structural precision
`may well become standard procedures in the near future.
`The turnover of HA is extremely rapid. It is estimated
`that of the 15 g of HA in the vertebrate body, 5 g turn over
`daily. The t1/2 of HA in the circulation is between 2 and
`5 min. In the epidermis of skin, where one half the HA of
`the body is found, it is one to two days, and in an
`apparently inert tissue as cartilage, it is approximately one
`to three weeks (Stern, 2003). It is evident that when HA is
`used for tissue scaffolding and matrix supports, modifi-
`cation of the HA is required to avoid the body's robust
`catabolic system. The various catabolic cascades and
`techniques are reviewed here. Though these reactions are
`
`useful in assembling the biomaterials, the same reactions
`must subsequently be avoided once the materials are
`inserted. Techniques to evade such reactions are the
`subject of a different review.
`
`2. Enzymatic catabolism of hyaluronan
`
`The hyaluronidases (Hyals) are a class of enzymes
`that degrade predominantly HA. The name is somewhat
`of a misnomer, since most of these enzymes have the
`limited ability to also degrade chondroitin (Ch) and
`chondroitin sulfates (CS). Both the prokaryotic and
`eukaryotic Hyals generally have the ability to degrade
`both classes of substrates,
`though a few bacterial
`enzymes have absolute specificity for HA.
`There is a major difference between the mechanisms
`that prokaryotic and eukaryotic enzymes utilize in cleaving
`their HA, Ch, and CS substrates. These differences are
`outlined below.
`
`2.1. Historical background
`
`identified as a
`Hyaluronidase activity was first
`“spreading factor” in extracts from mammalian testes
`(Duran-Reynals, 1928), and was also observed in bacterial
`extracts shortly thereafter (Duran-Reynals, 1933). The
`term “hyaluronidase” was then introduced to denote
`specifically the enzymes that degrade HA (Chain and
`Duthrie, 1940; Hobby et al., 1941). Karl Meyer classified
`these originally into three distinct categories, a scheme
`based on biochemical analyses of the enzyme reaction
`products (Meyer, 1971). With the advent of genetic data, it
`is now recognized that this classification scheme was
`remarkably accurate. Meyer was able to identify the three
`principal types of hyaluronidases, the two classes of
`eukaryotic endoglycosidase hydrolases described below,
`and the prokaryotic lyase-type of glycosidase.
`
`2.2. Prokaryotic enzymes that cleave HA
`
`2.2.1. Bacterial β-endoglycosidases and their lyase
`mechanism of action
`The bacterial enzymes are lyases that specifically cleave
`the β-(1→4) linkage in HA and CS. These are β-
`endoglycosidases that are eliminases with a mechanism of
`action entirely different from that of the eukaryotic
`glycoside hydrolases. These enzymes function by β-
`elimination with introduction of an unsaturated bond (EC
`4.2.99.1) (Jedrzejas, 2004; Stern and Jedrzejas, 2006). The
`generation of the double bond enables to perform a
`spectrophotometric assay, which is not available for the
`eukaryotic enzymes, which act hydrolytically. The relative
`
`

`

`540
`
`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`ease of determination of their activity has provided a wealth
`of information and thorough characterization (Jedrzejas,
`2004; Stern and Jedrzejas, 2006).
`The mechanism of action of these enzymes is of an
`acid/base processive type termed proton acceptance and
`donation (PAD). This is well established for the Strep-
`tococcal enzyme, but has not been shown conclusively
`for other bacterial enzymes. The catalytic mechanism
`involves several sequential steps (Li et al., 2000; Li and
`Jedrzejas, 2001; Ponnuraj and Jedrzejas, 2000):
`
`1. binding of the HA substrate to a cleft in the enzyme;
`2. acidification of C-5 carbon atom of a glucuronate
`residue by an enzyme Asn residue that functions as
`an electron sink;
`3. extraction of this C-5 carbon proton by an enzyme His
`residue, followed by the formation of an unsaturated
`bond between C-4 and C-5 of the glucuronate on the
`reducing side of the glycosidic bond;
`4. cleavage of the glycosidic bond after a proton has
`been donated from the enzyme Tyr residue;
`5. departure of the HA or CS disaccharide product from
`the active site and balancing of the hydrogen ions by
`an enzyme exchange with the water environment.
`
`The enzyme is then geared for another round of
`catalysis. For processive degradation of the substrate, HA
`is translocated by one disaccharide unit toward the re-
`ducing end of the chain and endolytically degraded using
`the PAD mechanism. During the process, the C-4 and C-5
`carbon atoms change their hybridization from sp3 to sp2
`with respective changes in the product conformation of
`the sugar ring, involving a puckering of the ring. This
`leads to a distorted half chair type of conformation.
`
`2.2.2. Bacterial β-exoglycosidases
`The exoglycosidases, including a β-glucuronidase
`and a β-N-acetyl-hexosaminidase are able to degrade
`HA, by removing single monosaccharide units. Endoly-
`tic cleavage creates new substrate sites for these
`exoglycosidases. However,
`the contribution to the
`overall catabolism of HA in bacteria or in their
`eukaryotic hosts is not known. The β-glucuronidase
`from E. coli is a ca. 290 kDa tetrameric protein with a pH
`optimum in the neutral range, in marked contrast with the
`acid-active enzyme from eukaryotes.
`
`2.3. Eukaryotic enzymes that cleave HA
`
`The eukaryotic enzymes that degrade HA have a
`surprisingly wide range of pH optima. These enzymes are
`hydrolases and mechanisms of their action have, until
`
`recently, defied thorough explication, largely because of
`the difficulties in assaying their activities.
`
`2.3.1. β-Endoglycosidases
`The eukaryotic endoglycosidase types of hyaluroni-
`dases were a long neglected class of enzymes (Kreil,
`1995). They are hydrolases, adding water across the
`cleavage site, and their activities are not detectable by
`spectrophotometry. These problems are now largely
`overcome, and much data are now accumulating rapidly,
`in part due to the human genome project and the EST
`(expressed sequence tag) data bank.
`
`2.3.1.1. Endo-β-n-acetylhexosaminidases. There are
`six such hyaluronidase-like sequences in the human
`genome (EC 3.2.1.35), but only three, of their products,
`Hyal1, Hyal2, and PH-20 are associated with known
`hyaluronidase enzymatic activities. The other proteins
`may have enzyme activities, but to date, the conven-
`tional in vitro assays have failed to detect them.
`They may have other activities, functioning perhaps as
`adhesion proteins. One of these proteins, Hyal2, acts as a
`cell-surface receptor for certain retroviruses (Rai et al.,
`2001). In fact, it can be suggested that all of these proteins
`have functions other than enzymatic activity.
`Three of the protein sequences Hyal2, Hyal4, and
`PH-20 are GPI-(glycophosphatidylinositol-) linked to
`outer cell-surface membranes. However, they also exist
`as free processed forms. Some of the processed forms
`have changes in pH optima (Oettl et al., 2003).
`Three sequences, Hyal1, Hyal2, and Hyal3 are located
`on chromosome 3 at 3p21.3, and another three, Hyal4,
`HyalP1 (a pseudogene, transcribed but not translated in
`the human), and SPAM-1 are clustered similarly on
`chromosome 7 at 7q31.3. Tetrasaccharides are the major
`catabolic products. The seventh sequence occurs in the
`mouse at the syntenic region, chromosome 6A2, termed
`rather unfortunately Hyal5 (since this is the seventh
`sequence to be identified, Hyal7 would be a more suitable
`appellation). The human genome may be the only
`mammalian one that lacks this sequence. The Hyal5
`protein is present in testicular extracts and may participate
`in penetration of the cumulus mass and in fertilization
`(Kim et al., 2005).
`The major hyaluronidase activities in somatic tissues
`are tentatively identified as Hyal1 and Hyal2 (Stern,
`2003, 2004). Hyal2 cleaves extracellular matrix HA to
`approximately 50 disaccharides (20 kDa), and then its
`activity slows down considerably. Hyal1, an acid-active
`lysosomal enzyme can accept HA of any size and
`rapidly cleaves these chains to small fragments. The
`tetrasaccharide is the predominant cleavage product.
`
`

`

`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`541
`
`These enzymes may work independently of each other
`in two separate pathways, or their activities may be
`coordinated. The relationship between the pathways is
`unknown. However, both are associated with CD44, the
`major HA receptor that can exist both extra- and
`intracellularly (Bourguignon et al., 2004; Harada and
`Takahashi, 2007).
`Based on high primary, secondary, and tertiary
`homology to the crystal structure of bee venom hyaluron-
`idase and site directed mutagenesis of the human PH-20
`enzyme,
`it
`is proposed that degradation by human
`hyaluronidase hydrolases may proceed via a double-
`displacement mechanism, with retention of HA substrate
`conformation. Such a mechanism involves one Glu amino
`acid residue as a hydrogen atom donor and a carbonyl
`oxygen of N-acetyl group of the HA performing the
`function usually assigned to a carboxyl group of another
`amino acid. The usual two-carboxylic-acids mechanism
`can be modified to reflect
`the different nucleophilic
`residues. The steps involved in this mechanism are:
`
`(1) binding of the HA substrate to the hyaluronidase;
`(2) residues around the catalytic site positioning the
`carbonyl oxygen nucleophile of the HA's N-acetyl
`group next to the to-be-cleaved β-(1→ 4) glyco-
`sidic bond, attacking the C-1 carbon of the same
`sugar to form a covalent intermediate between
`them. This leads to cleavage of the glycosidic bond
`and also results in the inversion of the anomeric C-1
`atom configuration;
`(3) at the same time a protonated Glu donates its H
`(deprotonation, acid function) to the glycosidic
`oxygen, leaving part of HA (the glycone part)
`unaltered;
`(4) hydrolytic cleavage of the intermediate bond
`between carbonyl oxygen and C-1 by a water
`molecule in the active site leads to re-protonation
`of a Glu, readying it for the next catalytic step, and
`to the second inversion of the configuration of C-1;
`(5) release of the HA product from the hyaluronida-
`se's active site (glycan on the aglycone side of the
`cleaved glycosidic bond).
`
`The anomeric configuration of the C-1 carbon atom of
`the substrate is retained throughout the process, being
`inverted twice during catalysis (in steps 2 and 4 above). The
`formation of an oxocarbonium ion transition state has been
`implicated in this process in step 2. Structural evidence from
`the enzyme-(HA tetrasaccharide) complex suggests that:
`
`(1) as the carbonyl nucleophile moves into place to
`interact with the C-1 carbon of HA, it results in
`
`change in puckering of the pyranose ring of N-
`acetyl-D-glucosamine on the non-reducing side of
`the bond to be cleaved, from regular chair to
`distorted boat and consequently
`(2) in moving the glycosidic bond into nearly an
`equatorial position, which results in its positioning
`closely to a Glu to allow for the donation of its H
`atom to this glycosidic oxygen as the bond is
`being cleaved.
`
`2.3.1.2. β-Endoglucuronidases. The hyaluronidases
`that are β-endoglucuronidases (EC 3.2.1.36) cleave the
`β-(1→ 3) glycosidic bond. No sequence data are available,
`thus it is not possible to compare enzymatic mechanisms,
`nor is it understood why the β-(1→3) is a more susceptible
`bond for cleavage than the β-(1→ 4) bond. They generate
`tetrasaccharides as the predominant catabolic end-pro-
`ducts, but also hexasaccharides. These enzymes, charac-
`teristic of annelids such as the leech Herudo medicinalis
`and certain crustaceans, utilize the hydrolysis mechanism.
`They thus resemble the vertebrate hyaluronidases more
`closely than the prokaryotic enzymes.
`
`2.3.2. β-Exoglycosidases
`There are two acid-active enzymes that remove single
`sugar units sequentially from the non-reducing termini of
`HA chains, both of which are lysosomal enzymes. The
`endolytic cleavage by the hyaluronidases generates
`increasing numbers of non-reducing termini as substrates
`for these exoglycosidases. The relative contribution of
`exo- and endoglycosidases to overall HA catabolism has
`not been established. The exoglycosidases appear to be in
`vast excess over the endoglycosidase enzymes. It is highly
`likely that such activities can contaminate impure prepa-
`rations of hyaluronidases, and generate spurious results.
`
`2.3.2.1. β-Exoglucuronidase. The activity of β-glucu-
`ronidase increases in many pathological conditions
`including liver inflammation, cholestatic jaundice, cirrho-
`sis of the liver, inflammations of other organs, tuberculosis,
`sarcoidosis and also in a number of neoplasms. Increased
`activity is also a sensitive indicator of aberrations in cell
`signaling networks.
`
`2.3.2.2. Exo-β-N-acetylglucosaminidase. The human
`protein has no apparent signal sequence, supporting the
`idea that this enzyme is localized in the cytosol. A gene
`database survey reveals the occurrence of enzyme homo-
`logues in the fruit fly, Drosophila melanogaster, in the
`nematode, Caenorhabditis elegans, and in plants indicates
`the broad occurrence of the enzyme in eukaryotes. The
`gene is expressed in the lysosomes of a variety of human
`
`

`

`542
`
`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`tissues, suggesting that it is involved in basic biological
`processes in eukaryotic cells, including the trimming of
`OSs in the cytosol of cells (Suzuki et al., 2002).
`
`2.4. Substrate specificity of the hyaluronidases
`
`The hyaluronidases of all origins do not have
`absolute substrate specificity, but have the ability to
`utilize Ch and CS as substrates as well. These two
`GAGs are far less effective substrates, and the reactions
`proceed more slowly (Rigden and Jedrzejas, 2003).
`There is a natural binding affinity between HA and CS
`(Turley and Roth, 1980). It may be that these two
`materials can exist in nature as a complex, and that the
`enzyme is adapted to degrading the two polymers
`simultaneously, at approximately a speed that reflects
`their relative abundance in nature.
`There is one class of prokaryotic hyaluronidases that
`does, however, have absolute specificity for HA, and
`these are the enzymes from Streptomyces hyalurolyticus
`(Ohya and Kaneko, 1970; Shimada and Matsumura,
`1980). This enzyme is entirely different from other
`bacterial hyaluronidases. It is used for structural studies,
`because of its absolute specificity. In many viscoelastics,
`CS and HA are combined, as in phaco-emulsifiers used
`in cataract surgery (e.g. Hutz et al., 1996). The specific
`and non-specific hyaluronidases may be useful comple-
`mentary tools for probing contributions of individual
`components of such preparations.
`
`loose association with albumin. Presentation of sub-
`strate to enzyme occurs most likely in the presence of
`such hyaladherins. How the enzyme is able to recognize,
`bind, and catabolize HA decorated with an array of
`varying proteins is a complete mystery.
`Albumin has long been identified as an HA-binding
`protein (Johnston, 1955; Niedermeier et al., 1966), as
`has fibrinogen (LeBoeuf et al., 1986). The presence of
`albumin, fibrin, or fibrinogen in preparations of matrices
`for tissue engineering most likely modulates enzymatic
`activity of hyaluronidases.
`c) Salt effects have been documented, which may
`result from a “salting out” of associated proteins, and
`their attendant effects. Albumin similarly can ameliorate
`salt effects on hyaluronidase activity, as has been shown
`for the liver-derived enzyme (Gold, 1982).
`d) Stability of hyaluronidase enzymes from eukaryotic
`sources is enhanced in the presence of detergents. The first
`somatic enzyme from a vertebrate source was dependent
`on the constant presence of detergents (Frost et al., 1997).
`However, it is well established that detergents can also
`function as protease inhibitors, and the detergent effect
`may be precisely that. The hyaluronidases in vertebrate
`tissues are potent enzymes that are present in exceedingly
`low concentrations. Protease inhibition may be the actual
`function of detergents on such highly active enzyme
`proteins present at extremely low concentrations. Recent
`isolation of Hyal1 was apparently conducted without
`detergents (Hofinger et al., 2007).
`
`2.5. Provisos for the hyaluronidase reactions
`
`2.6. The transglycosylation reaction
`
`There are a number of problems that will be
`encountered in working with the hyaluronidase
`enzymes, and in attempting to control their reactions.
`These are listed below:
`a) Effects of pH can be wide ranging. The pH optimum
`of the expressed human Hyal1 is pH 3.6, which is far more
`acidic than any cellular compartment. The pH of
`lysosomes is estimated to be pH 4.8. The in vitro assay
`reaction may be far different from the catalysis that takes
`place in vivo. Hyaluronidases undergo processing, and
`their apparent pH optimum can undergo changes in
`parallel. This is demonstrated for PH-20 (Oettl et al., 2003).
`b) Protein effects have not been thoroughly exam-
`ined. The characterization of the hyaluronidases is
`classically carried out in vitro utilizing highly purified
`enzymes and substrates. This is far removed from the
`situation in vivo. A wide range of HA-binding proteins,
`termed hyaladherins have been documented, ranging
`from tightly associated proteoglycan core proteins such
`as the amino terminal of aggrecan and link protein, to a
`
`An additional proviso for working with the hyalur-
`onidases is the mysterious transglycosylation reaction
`(Hoffman et al., 1956). This poorly understood reaction
`takes on intensity as the individual HA and CS chains
`decrease in size. It is not certain whether this reaction
`occurs in nature or whether it is an artifact of the reaction
`in vitro. It is capable of cross-linking individual HA
`chains, or forming hybrid chains composed of HA and
`CS. Neither has it been established whether such hybrid
`molecules exist
`in nature, nor whether they have
`biological activity. However, it must be kept in mind
`that such unwanted side-reactions may occur in preparing
`tissue matrices, in which transglycosylation may even
`predominate over the cleavage reaction.
`
`2.7. Inhibitors of hyaluronidases
`
`A further limitation for working with hyaluronidases,
`particularly in the presence of tissue-derived materials,
`is the ubiquitous presence of hyaluronidase inhibitors
`
`

`

`R. Stern et al. / Biotechnology Advances 25 (2007) 537–557
`
`543
`
`(Mio et al., 2000; Mio and Stern, 2002). These inhibitors
`are present in vast excess, as becomes obvious in the
`purification of these enzymes. The total units of enzyme
`activity increase after the initial steps in the purification,
`once inhibitors are separated from enzyme. Working
`with eukaryotic hyaluronidases in preparing tissue
`matrices must keep these provisos in mind.
`A number of compounds with intrinsic inhibitor
`activity have also been described (Mio and Stern, 2002).
`Many of these are derivatives of ascorbic acid (Botzki
`et al., 2004; Hofinger et al., 2007). These vitamin C-based
`inhibitors have the ability to inhibit bacterial as well as the
`vertebrate enzymes (Spickenreither et al., 2006).
`Heparin is an extraordinarily potent
`inhibitor of
`vertebrate hyaluronidases (Glick and Sylven, 1951),
`while the prokaryotic enzymes are impervious (R. Stern
`unpublished observation), as is the leech endoglucuroni-
`dase type of hyaluronidase (Jones and Sawyer, 1989).
`Other sulfated OSs are also effective inhibitors (Salmen et
`al., 2005). Among these are dextran sulfate (Lishanti et al.,
`2004), and probably sulfated proteoglycans, such as the
`syndecans (Tkachenko et al., 2005; Fears and Woods,
`2006). The opposing activities of HA and syndecans in
`many systems in suppressing and promoting differentia-
`tion, respectively, should be kept in mind when preparing
`materials for artificial matrices and tissue engineering.
`The presence of heparin and heparan sulphate, therefore,
`can have major effects on the ability to modulate HA-
`based tissue matrices using the various hyaluronidases.
`Finally, hyaluronidase action on HA results in
`formation only of even-numbered OSs having glucuro-
`nic acid at the non-reducing end and N-acetylglucosa-
`mine moiety at
`the reducing terminus (testicular
`hyaluronidase) or a reverse sequence of these two
`component monosaccharides (leech hyaluronidase)
`(Weissmann et al., 1954; Linker et al., 1960). Thus,
`preparation of the odd-numbered HA oligomers or even-
`numbered ones having an alternative structure would
`require additional
`treatment of
`the HA oligomers
`produced by hyaluronidase hydrolysis with chemical
`or enzymatic methods (Blundell and Almond, 2006).
`
`3. Non-enzymatic reactions that degrade HA
`
`3.1. Acidic and alkaline hydrolysis
`
`Similarly to other polysaccharides, HA can be
`degraded by acid or alkaline hydrolysis (Weissmann
`et al., 1953; Jeanloz and Jeanloz, 1964; Inoue and
`Nagasawa, 1985). However, chemical hydrolysis pro-
`ceeds in a random fashion and gives rise to a statistical
`mixture of oligo- and monosaccharides that can hardly be
`
`used for any specific purpose. Moreover, even short-term
`treatment of HA polymers at acidic or alkaline conditions
`can result in degradation, including “peeling” from the
`reducing end and β-elimination, characteristic for the
`uronic acid-containing poly- and OSs (Kiss, 1974).
`Similarly to other polysaccharides containing 4-substitut-
`ed or non-reducing glycuronic acids, HA under alkaline
`conditions degrades with the formation of unsaturated
`glucuronic acid units — 4-deoxy-hex-4-enoglucuronate
`(Yang and Montgomery, 2001). A “peeling” reaction
`occurs concomitantly, and leads to the formation of
`saccharinic acid (Whistler and BeMiller, 1958). At the
`same time, the N-acetylglucosamine component of HA is
`also degraded with alkaline treatment, giving rise to furan-
`containing species (BeMiller and Whistler, 1962). The
`authors warn that since HA is readily degraded in alkaline
`solutions, exposure to alkaline conditions during isolation
`should be avoided.
`In comparison to the effect of alkaline treatment,
`effects of lower pH conditions on HA have been much
`less investigated, and the results are even more
`ambiguous (Reed et al., 1989). No decrease of molar
`mass is observed at pH values below 2, and actu