Endoproteases of barley malt.pdf

18
Review Endoproteases of barley and malt Berne L. Jones * RR1, Box 6, Kooskia, ID 83539, USA Received 24 December 2004; revised 1 March 2005; accepted 1 March 2005 This review is dedicated to Dr Juhani Mikola, whose work laid the bases for many of our later studies on barley and malt endoproteases and their endogenous inhibitors. I never had the pleasure of meeting him, but if not for his early death, I expect that a very large percentage of the references in this article would have been to his work. Abstract During seed germination several seed biopolymers, including the storage proteins, must be hydrolysed to provide biochemical building blocks for the growing seedling. This process is particularly important in barley because under the guise of ‘malting’, it forms the basis of the malting and brewing industries. The steps involved in the enzymatic formation of ‘soluble protein’ during malting and in the ‘mashing’ phase of brewing are still not well understood. The barley proteins are initially solubilized by endoproteases and then further degraded by exopeptidases. The cysteine-class proteases probably play the most important roles, but their contributions are likely not as overwhelming as was thought previously. The metalloproteases are apparently also important players in protein solubilization, although their contributions have scarcely been examined. The characteristics of the purified aspartic class proteases imply that they are not important contributors to protein solubilization, but recent mashing studies indicate that they probably do play a minor role. All indications are that the barley and malt serine class proteases are not directly involved in storage protein hydrolysis during malting/mashing. More studies are needed to clarify the roles of the aspartic- and metalloproteases. One important aspect of further studies should be to ensure that appropriate biochemical methods are used, as well as conditions that are truly appropriate to commercial malting and mashing processes. q 2005 Elsevier Ltd. All rights reserved. Keywords: Endoproteases; Proteases; Barley; Malt; Mashing; Brewing; Inhibitors; Seed germination; Protease analysis 1. Introduction During and after the germination of barley seeds, many of the seed biopolymers must be broken down into their component subunits for use by the growing plant. One of the most important of these processes is the hydrolysis of proteins into peptides and amino acids. In addition to being a critical step for perpetuating species via seeds, this protein hydrolysis is critical to the brewing and malting industries, which are based on the preparation and use of malt. During the malting process, barley is germinated via a carefully controlled procedure so that its biopolymers, which cannot be utilized by yeast for growth and ethanol production, are degraded to sugars, amino acids and other low M r compounds which can be used in the brewing process. Much of the research in this field has been performed using various malting procedures, but the findings also apply to the general germination process. During malting, enough of the barley protein comp- lement must be degraded to amino acids and small peptides to provide sufficient nutrients for brewing yeasts to grow rapidly and to metabolize sugars into alcohol. Presumably, most of the proteins that are being degraded will be storage proteins, which are mainly hordeins. There is, however, no reason to think that various albumins and globulins are also not degraded during malting. The complete degradation of all of the barley proteins is not desirable because too little protein in beer (the main product made from malt) can result Journal of Cereal Science 42 (2005) 139–156 www.elsevier.com/locate/jnlabr/yjcrs 0733-5210/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2005.03.007 Abbreviations 2-D, two dimensional; ASBC, American Society of Brewing Chemists; 2-ME, 2-mercaptoethanol; E-64, (trans-epoxysuccinyl- L-leucylamido-(4-guanidino)butane); EP-A and -B, cysteine-class endo- proteases -A and -B; DTT, dithiothreitol; FAN, free amino nitrogen; HvAP, Hordeum vulgare aspartic protease; IEF, isoelectric focusing; kD, kilo Dalton; MEP-1, malt cysteine-class endopeptidase 1; MP, metalloprote- ases; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsul- fonyl fluoride; SEP-1, serine endopeptidase-1; SP, soluble protein. * Tel.: C1 208 926 4429. E-mail address: [email protected].

Transcript of Endoproteases of barley malt.pdf

  • Revie

    Endoproteases of b

    Berne L. J

    skia,

    d 1 M

    many

    th, I e

    o his

    Journal of Cereal ScienceE-mail address: [email protected]

    During seed germination several seed biopolymers, including the storage proteins, must be hydrolysed to provide biochemical building

    blocks for the growing seedling. This process is particularly important in barley because under the guise of malting, it forms the basis of the

    malting and brewing industries. The steps involved in the enzymatic formation of soluble protein during malting and in the mashing phase

    of brewing are still not well understood. The barley proteins are initially solubilized by endoproteases and then further degraded by

    exopeptidases. The cysteine-class proteases probably play the most important roles, but their contributions are likely not as overwhelming as

    was thought previously. The metalloproteases are apparently also important players in protein solubilization, although their contributions

    have scarcely been examined. The characteristics of the purified aspartic class proteases imply that they are not important contributors to

    protein solubilization, but recent mashing studies indicate that they probably do play a minor role. All indications are that the barley and malt

    serine class proteases are not directly involved in storage protein hydrolysis during malting/mashing. More studies are needed to clarify the

    roles of the aspartic- and metalloproteases. One important aspect of further studies should be to ensure that appropriate biochemical methods

    are used, as well as conditions that are truly appropriate to commercial malting and mashing processes.

    q 2005 Elsevier Ltd. All rights reserved.

    Keywords: Endoproteases; Proteases; Barley; Malt; Mashing; Brewing; Inhibitors; Seed germination; Protease analysis

    1. Introduction

    During and after the germination of barley seeds, many

    of the seed biopolymers must be broken down into their

    component subunits for use by the growing plant. One of the

    most important of these processes is the hydrolysis of

    proteins into peptides and amino acids. In addition to being

    a critical step for perpetuating species via seeds, this protein

    hydrolysis is critical to the brewing and malting industries,

    which are based on the preparation and use of malt. During

    the malting process, barley is germinated via a carefully

    controlled procedure so that its biopolymers, which cannot

    be utilized by yeast for growth and ethanol production, are

    degraded to sugars, amino acids and other low Mrcompounds which can be used in the brewing process.

    Much of the research in this field has been performed using

    various malting procedures, but the findings also apply to

    the general germination process.

    During malting, enough of the barley protein comp-

    lement must be degraded to amino acids and small peptides

    to provide sufficient nutrients for brewing yeasts to grow

    rapidly and to metabolize sugars into alcohol. Presumably,

    Abbreviations 2-D, two dimensional; ASBC, American Society of

    Brewing Chemists; 2-ME, 2-mercaptoethanol; E-64, (trans-epoxysuccinyl-

    L-leucylamido-(4-guanidino)butane); EP-A and -B, cysteine-class endo-

    proteases -A and -B; DTT, dithiothreitol; FAN, free amino nitrogen; HvAP,

    Hordeum vulgare aspartic protease; IEF, isoelectric focusing; kD, kiloRR1, Box 6, Koo

    Received 24 December 2004; revise

    This review is dedicated to Dr Juhani Mikola, whose work laid the bases for

    inhibitors. I never had the pleasure of meeting him, but if not for his early dea

    been tw

    arley and malt

    ones*

    ID 83539, USA

    arch 2005; accepted 1 March 2005

    of our later studies on barley and malt endoproteases and their endogenous

    xpect that a very large percentage of the references in this article would have

    work.

    42 (2005) 139156

    www.elsevier.com/locate/jnlabr/yjcrsproteins, which are mainly hordeins. There is, however, no

    reason to think that various albumins and globulins are also

    not degraded during malting. The complete degradation of

    all of the barley proteins is not desirable because too little

    protein in beer (the main product made from malt) can result0733-5210/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.

    doi:10.1016/j.jcs.2005.03.007

    fonyl fluoride; SEP-1, serine endopeptidase-1; SP, soluble protein.* Tel.: C1 208 926 4429.most of the proteins that are being degraded will be storageDalton; MEP-1, malt cysteine-class endopeptidase 1; MP, metalloprote-

    ases; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsul-

  • 2.2. FAN; free amino nitrogen

    until germination begins (steeping), is then held under moist

    catalytic mechanisms of these classes differ and the

    members of each are specifically inhibited by different

    eal SThe amount of NH2 groups in an extract, as measured

    by their interaction with a reagent that reacts specifically

    with this group. The reagent detects proteins, peptides and

    amino acids, but since the large proteins, intermediate

    peptides and small amino acids each contain only a single

    terminal NH2 group, a given weight of amino acid will be

    detected as having much more FAN than an equivalent

    weight of protein. FAN is therefore assumed to denote the

    amount of amino acids and small peptides in a wort. Since

    brewing yeasts can only metabolize these amino acids and

    small peptides, it is an indication of the nutritional value of

    an extract to the yeast.

    2.3. Mashing, extract, wort

    When ground malt is extracted with water whose

    temperature is increased in a carefully controlled manner,

    the industrial process is called mashing, and the productin a product that has insufficient foaming ability, mouthfeel

    and other required characteristics. The American Malting

    Barley Association, which is representative of the malting

    and brewing industries in the USA, currently suggests that

    between 42 and 47% of the protein of a commercially

    acceptable malt sample should be solubilised by the end of

    the mashing process (American Malting Barley Association,

    2004).

    Most of the research discussed in this review was

    conducted to obtain information about the biochemistry of

    malting so that the process of breeding improved malting

    barleys can be made more efficient and to enable maltsters

    and/or brewers to improve their processing methods.

    Some endoproteases from the leaves of barley plants

    have been studied in detail (Runeberg-Roos et al., 1994;

    Sundblom and Mikola, 1972; Thayer and Huffaker, 1984),

    but they will not be discussed because they do not directly

    affect grain or malt characteristics.

    2. Some definitions

    2.1. SP; soluble protein

    The amount of malt protein that is dissolved at the end of

    the mashing process. This value is determined either by

    measuring the nitrogen content of an extract or its UV

    absorbance and includes soluble proteins, peptides and

    amino acids. A mole of protein will contribute more to this

    value than will a mole of amino acid, since it will contain

    multiple nitrogen atoms and UV-absorbing groups. The SP

    of an extract is normally presumed to indicate its content of

    proteins and large peptides, which contribute to the foaming

    ability, mouthfeel and other physical properties of beers.

    B.L. Jones / Journal of Cer140is termed wort. The mashing processes used will vary,chemicals. For example, the cysteine proteases [EC

    3.4.22.-] contain the amino acid cysteine at their active

    centers and are specifically inhibited by a compound called

    E-64 (trans-epoxysuccinyl-L-leucylamido-(4-guanidino)bu-

    tane). The other protease classes and their commonly used

    specific inhibitors are the aspartic proteases [EC 3.4.23.-],

    inhibited by pepstatin A, serine proteases [EC 3.4.21.-],

    phenylmethylsulfonyl fluoride, or PMSF and metallopro-

    teases [EC 3.4.24.-], 1,10-phenanthroline, or o-phenanthro-

    line, and sometimes EDTA.

    Throughout this review the terms protease or endopro-

    tease have been used to cover the terms proteinase and

    endoproteinase when refering to enzymes that hydrolyse

    internal peptide bonds in proteins. The terms exopeptidase

    and peptidase will refer to enzymes that hydrolyse eitherand warm conditions for several days (germination) and

    finally is dried in a stream of air whose temperature is

    slowly raised (kilning). The method that is normally used at

    the USDA Cereal Crops Research Unit is described in Jones

    et al. (2000). When grain is germinated in this way, but not

    kilned, the product is called green malt. Malting

    conditions can be varied, depending on the malt character-

    istics needed. Barley grain that is germinated on filter paper

    or some other medium and air-dried or freezedried is not

    malt, but is simply germinated barley. It should be

    remembered that the biological germination of barley

    begins during the steeping process and is already well

    established when what maltsters call germination begins.

    2.5. Class-specific proteases

    Most endoproteases fall into one of four classes. Thedepending on what attributes are wanted in the final wort

    and beer. When the mashing is performed experimentally to

    measure the characteristics of a malt sample, using

    conditions that, in the USA, are strictly specified by the

    American Society of Brewing Chemists (American Society

    of Brewing Chemists, 1992, method Malt-4), the final

    product is called an extract. The process conditions used to

    produce extracts and worts differ significantly, although the

    terms are sometimes used interchangeably. Generally, the

    initial mash temperature is held constant for a time at a

    relatively low level, and this step is called a protein rest.

    The mash temperature is normally then increased at a

    constant rate (ramping) and then held constant, usually at

    around 70 8C (conversion), until all of the starch has beenhydrolysed to fermentable sugars.

    2.4. Malting

    The process of preparing malt. Barley grain is soaked

    cience 42 (2005) 139156the N- or C-terminal peptide bonds of polypeptides.

  • real S4. Meaningful protease assays

    The protease activities of barley and malt have been

    measured in many ways in the past, and the results obtained

    from a given study can vary greatly, depending on how the

    measurements are made. This makes it imperative that, to

    get meaningful results, appropriate analytical methods and

    conditions must be used. This has not always been the case.

    Some of the methods used were: measuring the amino

    acids released during mashing (Jones and Pierce, 1967a,b)3. Early studies

    Most studies performed on the proteolytic activities of

    barley and malt prior to 1980 are not discussed in detail,

    because they used preparations that were poorly character-

    ized and contained complex mixtures of proteases. These

    studies were, however, quite important, because they laid a

    strong foundation for further, more specific studies. For

    example, they showed that the proteolytic activity of mature

    barley grain was low, but that during malting the activity

    increased greatly (Burger and Prentice, 1970; Enari et al.,

    1964; Kringstad and Kilhovd, 1957). It was also shown that

    barley, and especially malt, contained several proteases

    (Burger, 1973; Burger and Prentice, 1970; Enari et al., 1964;

    Enari and Mikola, 1968) although the number of malt

    proteases was still underestimated. These studies indicated

    that the solubilization of barley proteins to SP was catalyzed

    by the endoproteases, and that the SP polypeptides were

    then further hydrolysed by exopeptidases to yield FAN

    (Mikola, 1983). It appeared that the cysteine class proteases

    catalyzed most of the SP release (Burger, 1973; Enari et al.,

    1964; Enari and Mikola, 1968; Sundblom and Mikola,

    1972), but that aspartic- (Morris et al., 1985) and

    metalloproteases (Enari et al., 1964; Enari and Mikola,

    1968; Sundblom and Mikola, 1972) were also probably

    involved. It was found that SP was released during both

    malting and mashing, although the contributions of each of

    these procedures to the final mash SP values varied

    considerably, depending on the processing and analytical

    methods used (Barrett and Kirsop, 1971; Burger and

    Schroeder, 1976a).2.6. 2-D IEF!PAGE

    A separation method in which the components of extracts

    are resolved by isoelectric focusing (IEF), after which the

    IEF gels are incorporated at the top of non-denaturing

    polyacrylamide gel electrophoresis (PAGE) gel slabs so that

    PAGE separations can be performed in the second

    dimension (Zhang and Jones, 1995a). To detect proteolytic

    activities, the second dimension (PAGE) gel normally

    contains a proteinase substrate, most often gelatin.

    B.L. Jones / Journal of Ceor from a hordein preparation (Baxter, 1976; Phillips andWallace, 1989), measuring the reduction in viscosity of a

    gelatin solution (ten Hoopen, 1968), and following the

    hydrolysis of either radioactively-labeled proteins (Morris

    et al., 1985), synthetic peptides (Suolinna et al., 1965),

    derivatized amino acids (Galleschi and Andreoni, 1990), or

    even the protein layer on a photographic film (Burger and

    Schroeder, 1976b).

    Some questions need to be asked about these and similar

    methods to ensure that they gave meaningful and relevant

    results. Among these are:

    (1) Was the right endoproteinase substrate used? For

    years researchers have been seeking easier and faster ways

    to measure endoproteinase activities. One of the most

    common has been to use a small peptide or a peptide

    analogue as a substrate instead of a protein. These methods

    have the advantage that the hydrolysis products are easy to

    quantify, compared to the polypeptide mixtures that are

    produced from proteins. However, the results obtained using

    these unnatural substrates may not reflect the actual

    reactions that occur when the enzymes hydrolyse proteins.

    For example, Jones and Poulle (1990) characterized the

    hydrolytic specificity of a purified 30,000 Mr green malt

    endoproteinase using as substrates two small purified

    proteins with known amino acid sequences that occur

    naturally in barley. In 1989 Phillips and Wallace isolated a

    very similar or identical proteinase, and determined its

    specificity using low Mr amino acid esters. The specificities

    obtained by these two methods were very different, although

    the enzymes appeared to be quite similar otherwise. The

    discrepancy was most likely due to the fact that the low Mrsubstrates, because of their very small sizes, did not contain

    the amino acid residues that really defined the specificity of

    the proteinase. It seems obvious that the best experimental

    substrate should be the one whose structure is most similar

    to that of the enzymes natural substrate; that is, a protein.

    But what protein substrate should be used? Not all

    proteins are equally useful for measuring proteolytic

    activities. This is especially true in the case of malt,

    where multiple proteases are present whose specificities

    differ. Thus azocasein, for example, is a poor substrate for

    green malt proteinase extracts (Phillips and Wallace, 1989)

    and we have found that when malt proteases are separated

    by 2-dimensional (2-D) electrophoresis only a few of the

    separated enzymes hydrolysed either azocasein or haemo-

    globin, two commonly used substrate proteins. Thus they

    are not appropriate for measuring the multiple activities of

    barley and malt. We have found that the best substrates for

    measuring malt proteases are gelatin or its colored

    derivative, azogelatin. These proteins are readily hydrolysed

    by serine-, cysteine- and metalloproteases from malt and

    other sources. They are more slowly degraded by most

    aspartic class proteases (Jones et al., 1998), including those

    of malt. These latter enzymes are best analyzed using the

    substrate edestin. A major advantage of using gelatin or

    azogelatin substrates is that they are soluble at pH values

    cience 42 (2005) 139156 141ranging from 3.0 to 10.5, and thus can be used to measure all

  • eal Sof the various malt proteases (other than the aspartic

    proteases) whose pH optima cover this entire pH range.

    Most of the proteins that have previously been used as

    substrates are not soluble throughout this extended pH

    range.

    (2) When substrates are used in different physical forms,

    how does this affect the hydrolysis results? For example,

    several studies have been performed with the substrate

    gelatin in its soluble form (ten Hoopen, 1968; Jones et al.,

    1998), as a layer on a plastic backing (Burger and

    Schroeder, 1976b), and as, most likely, a suspension of

    molecules entrained in an acrylamide matrix (Zhang and

    Jones, 1995a). Little or nothing is known about how these

    different forms of gelatin affect its susceptibility to

    hydrolysis. We have performed hydrolyses using both

    gelatin and azogelatin substrates in solution and also

    entrapped inside acrylamide gels and have occasionally

    observed differences in the hydrolytic actions of malt

    proteases on these substrates (Jones and Budde, 2003).

    (3) Can natural proteins be used as substrates? Because

    barley storage proteins (hordeins) are the major proteins

    hydrolysed during malting and mashing, it would be helpful

    to be able to use these proteins as substrates in enzyme

    characterizations. However, by definition hordeins are

    insoluble in both water and dilute salt solutions, conditions

    that are normally used for enzyme studies. The polymeric

    hordeins (hordenins) are even less soluble under these

    conditions. Thus for hordeins to be used as substrates in

    vitro, they must either be rendered soluble or the substrate

    must be present as a suspension, rather than in solution. To

    solubilize hordein fractions requires very harsh treatments

    and after solubilization the proteins are, by definition, no

    longer hordeins and thus are no longer truly natural

    substrates. Even if the hordeins are simply extracted from

    the grain and used as a suspension the extraction process is

    still quite vigorous and then the problem arises of whether

    the suspended solid hordein is hydrolysed differently from

    hordein in solution.

    When protein solubilization occurs during malting and

    mashing, the initial hordein hydrolysis presumably involves

    a depolymerisation of the proteins that are still in their

    native states. As the depolymerisation continues (as the

    grain is modified) and the internal structure of the grain is

    lost, the environment in which the hordeins reside changes,

    and their molecular structures may also be modified.

    Buchanan and Kobrehel and their collaborators (Besse

    et al., 1996) have proposed that during grain modification

    some of the disulphide bonds in hordein may become

    reduced, rendering the proteins more soluble. However, gel

    electrophoresis has indicated that the hordeins remain

    relatively unchanged during malting, with the individual

    protein bands simply becoming fainter and fainter as

    malting proceeds (Marchylo and Kruger, 1985; Poulle and

    Jones, 1988; Smith and Simpson, 1983). However to

    perform these electrophoretic analyses the hordein samples

    B.L. Jones / Journal of Cer142had first to be solubilized, which probably also modifiedtheir structures. Although the hordein extracts used as

    endoproteinase substrates do not really contain native

    hordein molecules, presumably their amino acid sequences

    are unaltered, although their secondary, tertiary and/or

    quaternary structures have probably been changed. In any

    case, when hordein suspensions have been used as

    substrates some purified proteases hydrolysed them and

    others did not, and this has been used as a major criterion of

    whether a given enzyme is involved in solubilizing storage

    proteins in vivo.

    (4) Have characterizations been performed at the

    appropriate pH values? It has been noted since the earliest

    studies that the different barley/malt endoproteases showed

    maximal activities at different pH values, with the greatest

    activity of extracted enzyme mixtures occurring at low pH

    levels. Subsequently, many proteases have been studied at

    pH values between 3 and 4. However, the pH of the

    endosperm of germinating barley is about 4.9 (Mikola and

    Virtanen, 1980), and that of a North American mash is

    around 5.86.0 (Jones and Budde, 2003), so many of the

    data from those early studies are not really relevant to the

    events that occur during either malting or mashing (see

    Section 7.3). A similar problem has arisen due to the adding

    of reducing agents to enzyme extracts and to their analysis

    mixtures. It has been shown recently that the enzymatic

    hydrolysis of protein substrates by malt enzymes is strongly

    enhanced in the presence of reducing agents and lowered by

    oxidizing agents (Jones and Budde, 2003). Nearly all earlier

    barley/malt proteinase analyses were performed in the

    presence of added reducing agents, so many of the results

    are probably not indicative of what really occurs during the

    processing of barley and malts. Specific examples of these

    problems are discussed in Section 7.4. To ensure that their

    results are relevant, researchers hoping to apply their

    findings to real systems need to be aware of how those

    systems operate and to ensure that they use appropriate

    conditions and techniques.

    (5) Are meaningful data being collected and reported?

    One of the most basic aspects of enzymology involves the

    measurement of initial rates of enzyme catalyzed reactions,

    and holds that this measurement should be made, if at all

    possible, while the reaction rate is still linear. To ensure this,

    multiple measurements of the concentrations of either the

    reactants or products must be made and it must be shown

    that these reaction components are utilized or released at a

    constant rate throughout the measurement period. Unfortu-

    nately, this basic enzymological principle is often over-

    looked, and the cereal endoproteinase literature is replete

    with reports of inappropriately measured activities. The

    hydrolysis rates of most endoproteinase-catalyzed reactions

    performed in solution are generally linear for 30 min or less

    (Jones et al., 1998), so any data obtained using a single

    measurement made several hours after a reaction has been

    started, without previously proving that the rate is constant

    cience 42 (2005) 139156throughout that period, are at best only semi-quantitative.

  • real S5. Purified barley/malt endoproteases

    Between about 1985 and 1990, there was a burst of

    publications that described the purification and character-

    ization of individual proteases from barley tissues or malt.

    5.1. Aleurain

    In 1985, Rogers and co-workers (Rogers et al., 1985)

    isolated and sequenced a cDNA clone from gibberellin-

    treated barley aleurone cells that apparently encoded a 361-

    amino acid protein. Because of the similarity of its amino

    acid sequence with those of cathepsin H and two plant thiol

    proteases, they concluded that the protein was a thiol

    endoproteinase, and named it aleurain. It was speculated

    that it might be the proteinase that was being studied

    concurrently by Hammerton and Ho (1986). However, when

    barley leaf aleurain was finally purified and characterized, it

    was shown to be an aminopeptidase, rather than an

    endoproteinase (Holwerda and Rogers, 1992), so it is

    unlikely to play any major part in solubilizing grain storage

    proteins. This is but one of many examples of a

    phenomenon that researchers too often forget: that although

    amino acid or DNA base sequence homologies can indicate

    possible biochemical similarities, until the protein is

    isolated and characterized, nothing is proven.

    5.2. Cysteine-class proteases

    5.2.1. EP-A and EP-B

    Hammerton and Ho (1986) showed that gibberellic acid-

    treated barley aleurone layers synthesized several proteases

    with Mr of w37,000, that were inhibited by cysteineprotease inhibitors, and that hydrolysed extracted barley

    storage proteins. Two of these enzymes, called endopro-

    teases-A (EP-A) (Koehler and Ho, 1988) and -B (EP-B)

    (Koehler and Ho, 1990a) were purified. The EP-A

    preparation had a Mr of 37,000, hydrolysed internal peptide

    bonds in substrate proteins, had a pH optimum of 5, and

    contained three very similar isozymes. The enzyme(s) was

    not aleurain, but was apparently closely related to papain, a

    papaya (Carica papaya) proteinase. The purified EP-B had a

    Mr of 30,000, contained two proteins with pI values of 4.6

    and 4.7, and behaved as a cysteine endoproteinase on the

    substrate haemoglobin. Its N-terminal amino acid sequence

    and properties were similar to those of EP-A, and both EP-A

    and -B produced similar polypeptides when they hydrolysed

    a hordein preparation (Koehler and Ho, 1990a).

    The cDNA cloning of EP-B has been performed and

    reported (Koehler and Ho, 1990b). Three clones were

    obtained, two of which encoded EP-B isozymes that

    were 98% similar and showed large preprosequences.

    These were converted via a multistep process into the

    mature enzymes. The processing of proEP-A was less

    complicated and the final form of the enzyme was secreted

    B.L. Jones / Journal of Cefrom the aleurone tissue (Koehler and Ho, 1990b).Gibberellin stimulated the aleurone to produce both EP-A

    and -B, but EP-B was synthesized more quickly than EP-A.

    Not surprisingly, the mRNA for EP-B was essentially absent

    from mature barley grain, but increased strongly within 1 or

    2 days of germination. Initially, the EP-B was expressed in

    the scutellar epithelium and aleurone cells adjacent to the

    embryo. Later the mRNA concentration was highest in the

    aleurone layer adjacent to the endosperm, and, with time, its

    concentration increased along the length of the grain to its

    distal end (Marttila et al., 1993). The concentration of newly

    synthesized EP-B protein showed a similar distribution.

    These findings are all consistent with endoproteinase EP-B

    being one of the major enzymes responsible for degrading

    barley storage proteins during seed germination. Mikkonen

    et al. (1996) found evidence for only two EP-B genes in

    barley, both residing on chromosome 3. The hormonal

    regulation of one of the barley EP-B genes has since been

    described (Cercos et al., 1999).

    5.2.2. Malt endopeptidase 1 (MEP-1)

    Phillips and Wallace, working at the same time as

    Koehler and Ho, but using green (unkilned) malt, found that

    the predominant proteinase activity under their assay

    conditions, was also a cysteine class protease (Phillips and

    Wallace, 1989). They purified and characterized the

    enzyme, named it MEP-1, and showed that it hydrolysed

    hordein in suspension. The hordein had been solubilized

    with 50% alcohol and, as discussed in Section 4, was

    presumably not physically the same as native hordein. In

    addition, only a single reaction sample, taken after 1 h of

    incubation, was analyzed, so it seems likely that the values

    reported are not true initial reaction rates. The enzyme also

    hydrolysed azocasein and haemoglobin, but again only

    single 2 h reaction samples were analyzed. The purified

    MEP-1 migrated on SDS-PAGE as a single band of Mr29,000, but on isoelectric focusing it separated into two

    components with pI values of 4.2 and 4.3 (Phillips and

    Wallace, 1989).

    When MEP-1 hydrolysed a series of N-t-butoxycarbonyl-

    L-amino acid-p-nitrophenyl ester substrates containing

    various amino acid residues, the hydrolysis rates were:

    GlnOAlaOLeuOTyrOTrpOAsnOPhe. The hydrolysis ofhordein was strongly increased by the addition of the

    reducing agent 2-mercaptoethanol (2-ME), whereas the

    hydrolysis of azocasein and of the synthetic arginine

    substrate were only weakly increased. An antibody raised

    against MEP-1 also cross-reacted with a Mr 37,000

    endoproteinase. It was later reported (Guerin et al., 1992)

    that the amino acid sequence of the first 20 residues of MEP-

    1 was identical with that of the EP-B studied by Ho and his

    collaborators, and that the gene encoding MEP-1 was

    located on the long arm of chromosome 3 of the barley cv

    Betzes.

    Although there are some discrepancies between

    Wallaces enzymes and those studied by Ho, it seems

    cience 42 (2005) 139156 143obvious that MEP-1 and EP-B are the same, or very closely

  • cal to that found with the 30 kD protease, except that Ala did

    eal Srelated, proteases. Among the more important outcomes of

    Wallaces study were the findings that MEP-1 (EP-B) and,

    probably, EP-A were present in germinating barley, as well

    as in the isolated aleurone tissue, and that the purified green

    malt enzyme could hydrolyse extracted barley storage

    proteins.

    5.2.3. The 30 kD proteinase of green malt.

    At the time the EP-B and MEP-1 enzymes were being

    studied, Jones and Poulle independently purified and

    characterized a proteinase from green malt that they

    designated the 30 kD endoproteinase. The native 30 kD

    enzyme had an apparent Mr of 20,000 from gel filtration

    analysis, but SDS-PAGE indicated that it was a single

    polypeptide of Mr 30,000 (Poulle and Jones, 1988). It had a

    pH optimum of 3.8 and was a cysteine class enzyme that

    hydrolysed several large proteins. Its amino acid compo-

    sition was quite different from that of aleurain.

    The purified enzyme rapidly hydrolysed all three barley

    hordein classes (B, C and D), with the B and D hordeins

    being degraded faster than the C proteins (Poulle and Jones,

    1988). When a mixture of B and D hordeins prepared by

    HPLC was used as substrate, a few discrete, intermediate

    sized peptide reaction products were detected by PAGE, but

    most of the hordein was hydrolysed to peptides that were too

    small to be retained on the acrylamide gel. The changes that

    occurred in the SDS-PAGE patterns of a hordein prep-

    aration that was hydrolysed by the purified 30 kD protease

    were very similar to those seen in hordein samples that were

    removed from barley undergoing malting, indicating that

    the purified enzyme was behaving like those that were

    active during malting. Immunomicroscopy of aleurone

    tissues (Marttila et al., 1995) showed that after synthesis

    this enzyme was transported to the starchy endosperm,

    where the storage proteins are located. From the similarities

    shared by the 30 kD, EP-B and MEP-1 enzymes, it seems

    apparent that they were the same endoproteinase.

    5.2.4. The 31 kD proteinase of green malt.

    Zhang and Jones (1996) purified and characterized a

    second endoproteinase from 4-day germinated green malt.

    It was also a cysteine protease, with a pI of 4.4, was

    hydrolytically most active at pH 4.5, and its SDS-PAGE

    molecular mass was about 31,000. It hydrolysed gelatin,

    azocasein, haemoglobin, edestin and a hordein prep-

    aration. The 31 kD proteinase was one of five proteins in a

    crude malt extract that reacted with antibodies raised

    against the 30 kD proteinase of Poulle and Jones (1988).

    Even though they were well separated on non-denaturing

    Western blot acrylamide gels, all five cross-reacting

    proteins migrated to the Mr 30,00031,000 zone of gels

    that contained SDS and reducing agent (Zhang and Jones,

    1996). The sequence of the N-terminal nine amino acids

    of the 31 kD protein was identical to that of one of the Mr37,000 EP-A isozymes studied by Koehler and Ho (1988),

    B.L. Jones / Journal of Cer144but the 31 kD enzyme did not cross-react with antibodiesnot specify hydrolysis with the 30 kD enzyme. This order is

    very nearly the same as that of the hydrophobicities of the

    amino acids at pH 7 (Creighton, 1984). As with the 30 kD

    enzyme, the presence of the large hydrophobic amino acid

    pyridylethylcysteine at the P2 site was not enough to

    promote hydrolysis, but hydrolysis did occur when two suchraised against EP-A. Conversely, its N-terminal amino

    acid sequence differed from that of the Mr 30,000 EP-B,

    but it did cross-react with antibodies raised against the

    30 kD malt proteinase of Poulle and Jones (1988), which

    was very similar to EP-B.

    5.2.5. Hydrolytic specificities of the cysteine proteases

    To evaluate its specificity, purified 30 kD green malt

    proteinase was used to hydrolyse two small barley proteins,

    the a- and b-hordothionins, into peptides (Jones and Poulle,1990). As the native forms of the hordothionins were

    intractable to hydrolysis by the 30 kD enzyme and other

    proteases, they were reduced and alkylated prior to use.

    Hydrolyses were performed for varying periods and

    analyses of the resultant peptides defined the exact bonds

    hydrolysed by the protease and their relative rates of

    hydrolysis. It was apparent that the principal specificity of

    the 30 kD enzyme was not defined by either of the amino

    acids directly involved in forming the hydrolysed peptide

    bond. The enzyme specifically cleaved hordothionin peptide

    bonds between the P1 and P01 residues of hydrolysis sites

    that had the general formula

    NH2 /P2 KP1YKP01 KP

    02 /COOH

    when the amino acid residue located at the P2 site was either

    Leu, Val or Tyr, with Leu specifying the fastest hydrolysis.

    Because hordothionin substrates contained no Ile or Trp,

    and the single Phe was located at the C-terminus, and thus

    not available for endoproteolytic hydrolysis, the effects of

    these other large aliphatic or aromatic residues could not be

    tested. Although the amino acid in the P2 site was the major

    factor for determining hydrolysis, the residues occupying

    the P1 and/or P01 sites also seemed to have some small effect

    on the hydrolyses (Jones and Poulle, 1990).

    The specificity of the 31 kD malt proteinase was

    determined similarly, using as substrates reduced and

    alkylated b-purothionin and several other polypeptidesthat were chosen to ensure that one or more of them

    contained each of the 20 common amino acids (Zhang and

    Jones, 1996). b-Purothionin is a wheat protein that ishomologous to the barley hordothionins. Analysis of the

    resultant peptides showed that the specificity of the 31 kD

    enzyme was similar to that of the 30 kD enzyme; hydrolysis

    was determined mainly by the amino acid occupying the P2position relative to the bond that was hydrolysed. The

    effectiveness of amino acids for specifying hydrolysis was

    TrpOPheOLeuOIleOValOTyrOAla, which was identi-

    cience 42 (2005) 139156residues were located together at positions P2 and P3.

  • real SDavy et al. (1998) confirmed these results by hydrolysing

    a recombinant C hordein molecule with EP-B (Mr 30,000),

    and a group of synthetic peptide derivatives with both EP-A

    and EP-B. The hydrolytic sites were primarily specified by

    the amino acids at the P2 position, in the order LeuOPheOVal[ProOSer. They also found that hydrolysis wasseverely restricted when Pro was present at either the P1or the P01 sites. This imposes a rather strict limitation on thenumber of sites on barley storage proteins that are available

    for hydrolysis, since they are all proline-rich proteins. Some

    secondary hydrolysis sites were detected, but it was not

    clear what specified hydrolysis at these locations. In a

    second study using a larger and more directed set of

    synthetic peptides, Davy et al. (2000) again found that the

    major specifying site was at residue P2 for both EP-A and

    EP-B. The amino acids listed above for the 31 kD enzyme

    were most effective, except that Ala was replaced by Met.

    Specificity due to Met might have been overlooked in the

    study of the 31 kD enzyme (Zhang and Jones, 1996) because

    it was not present in the studied peptides at locations where

    its effect would have been readily discerned. The hydrolysis

    of peptides that were designed to highlight any specificity

    that was due to the amino acids at the substrate P1 site

    showed that the residue at this position had only a small

    effect on the hydrolysis rates, except that when Pro was

    present, no hydrolysis occurred.

    When the hydrolyses of wild type and mutated forms of a

    recombinant hordein C molecule by EP-B were compared,

    the hydrolysis of the wild type protein was again determined

    by the residues at the P2 site, while the hydrolysis of the

    mutated protein molecule, whose P2 specifying sites had

    been removed, occurred at a very slow rate (Davy et al.,

    2000). One important finding from this study was that the

    hordeins that were present in isolated protein bodies were

    hydrolysed by the EP-B, adding credence to the hypothesis

    that the malt cysteine proteases really do hydrolyse storage

    proteins in vivo.

    5.2.6. Cysteine proteinase summary

    From a comparison of their characteristics, it appears that

    the aleurone-derived proteinase EP-B is probably identical

    with the MEP-1 and 30 kD proteases isolated from green

    malt. The 31 kD protease apparently differs from both EP-A

    and EP-B, but is quite similar to both. The 30 and 31 kD

    proteases are probably isoenzymes, and it seems likely that

    malt contains at least three other very similar enzymes. All

    of these purified endoproteases were reported as being

    major activities in either barley or malt, presumably being

    the main proteases responsible for degrading storage

    proteins. However, this is not necessarily the case, because

    in essentially all cases the enzyme purifications and activity

    assays were performed in the presence of added reducing

    agent, usually 2-ME. As is discussed in Section 7.4, the

    addition of reducing agents to these particular enzymes

    greatly increases their activities. Thus, the 2-ME addition

    B.L. Jones / Journal of Cewould have enhanced their activities relative to those ofother enzymes that were present and thereby led to the

    conclusion that they play a bigger role in vivo than in fact

    they do. In addition, most of the enzyme assays were

    performed at pH values lower than the pH 4.85.0 of

    germinating barley seeds (Henson, C., personal communi-

    cation; Mikola and Virtanen, 1980). Because the cysteine

    proteases are most active at these low pH values, whereas

    the metallo- and serine endoproteases are most active at

    higher pH values, this would also make it appear that the

    cysteine proteases were relatively more important to seed

    germination than they really are. It is, however, probably

    significant that the specificities of these enzymes are

    admirably adapted for quickly digesting storage proteins

    to relatively small peptides. These can then serve as

    substrates for the exopeptidases that release the amino

    acids that are required by germinating seeds and by brewers

    (Zhang and Jones, 1996).

    5.3. Aspartic endoproteases

    5.3.1. HvAP

    Doi et al. (1980) showed that dormant rice seeds

    contained an acid protease that was inhibited by pepstatin,

    indicating that it was an aspartic proteinase, and Belozersky

    et al. (1989) purified and characterized aspartic proteases

    from seeds of wheat and buckwheat (Fagopyrum esculen-

    tum), a pseudo-cereal (Belozerskii et al., 1984). This led

    Sarkkinen et al. (1992) to purify a similar enzyme from

    barley seeds. Their preparation contained two proteases; one

    was a precursor molecule of Mr 48,000, the other the mature

    enzyme of Mr 40,000, each comprising two subunits. The

    enzyme was most active at pH around 3.7 and its activity

    declined quickly above pH 4.0. It did not hydrolyse either a

    barley globulin or a reduced and alkylated purothionin, but

    did degrade haemoglobin. Apparently, the ability of the

    enzyme to hydrolyse barley storage proteins was not tested.

    Incubation with class-selective inhibitors indicated that it

    was an aspartic proteinase.

    The cloned and sequenced cDNA encoding the enzyme

    showed clearly that the two enzyme forms were translated

    as a single proenzyme and then processed into the two

    forms, the Mr 40,000 form being the final product

    (Runeberg-Roos et al., 1991; Tormakangas et al., 1991).

    The enzymes amino acid sequence and inhibition charac-

    teristics were similar to those of mammalian and yeast

    aspartic proteases, especially cathepsin D. The enzyme was

    named Hordeum vulgare Aspartic Proteinase (HvAP) and

    its specificity was determined using small peptide substrates

    (Kervinen et al., 1993). Hydrolysis was maximal when the

    substrate contained either aromatic or aliphatic amino acids

    in both its P1 and P01 sites. Vacuoles of barley leaves and

    roots have an enzyme that is very similar to HvAP that

    processes a prolectin (Runeberg-Roos et al., 1994), but it is

    not clear whether this enzyme is the same as the one in

    seeds. Expression of the expression of HvAP in germinating

    cience 42 (2005) 139156 145and developing barley seeds (Marttila et al., 1995;

  • eal STormakangas et al., 1994) showed that the enzyme was

    present in aleurone cells but was not secreted into the

    starchy endosperm, where most storage proteins reside.

    Based on its characteristics, it is clear that HvAP is not

    involved in the solubilization of barley storage proteins,

    although it would be of interest to know whether the purified

    enzyme hydrolyses barley hordeins.

    5.3.2. Aspartic proteases in dormant and germinating

    barley seeds

    Using 2-D electrophoresis, Zhang and Jones showed that

    there were several aspartic proteases in germinated barley

    (Zhang and Jones, 1995a,b. Section 6.2), and it seemed

    likely from their characteristics, their seed locations, and the

    changes that occurred during malting that they might play

    an important role during germination. This supposition was

    tested by comparing the aspartic proteases of barley seeds

    and green malt (Zhang and Jones, 1999). Both seeds and

    green malt contained multiple aspartic protease forms, four

    in malt and six in seeds. These different forms were

    separated by pepstatin A affinity chromatography, followed

    by 1-D and 2-D IEF and gel electrophoresis. Three of the

    four separated malt proteases cross-reacted with antibodies

    raised against HvAP. All of the seed and malt proteases

    digested edestin, a Cannabis sativa globulin, most actively

    at pH 3.54.5 and of all the class-specific inhibitors, only

    pepstatin A caused inhibition. The enzymes all had identical

    pI values, except for one of the seed forms. SDS-PAGE

    showed that the seed proteases had subunit sizes of Mr 8000;

    18,000; 31,000; 38,000 and 48,000 and the green malt

    subunits were Mr 8000; 11,000; 15,000; 18,000; 31,000 and

    38,000. Thus, the barley and malt enzymes were very

    similar to each other and to HvAP, and the multiple

    forms probably arose by post-translational processing, as

    occurs with HvAP. None of the proteases hydrolysed

    the components of a hordein preparation, but they did

    associate with the barley chloroformmethanol soluble or

    CM proteins (Shewry, 1993) during the purification

    process and they digested some of the components of a

    CM protein mixture. Thus this study reinforces the previous

    findings for HvAP, indicating that these particular enzymes

    are apparently not involved in storage protein solubilization

    during malting (see also Section 7.5).

    5.4. Barley metalloproteases

    In some of their earliest studies Enari and Mikola

    (1968) reported that metalloproteases, as well as cysteine

    proteases, were present in barley seeds and in kilned and

    green malts. Later they showed that an EDTA-inhibited

    protease was synthesized and secreted from GA3-stimul-

    ated barley aleurone cells (Sundblom and Mikola, 1972).

    Also, Belozersky, Dunaevsky and their associates found a

    metalloprotease in dormant buckwheat seeds (Voskoboi-

    nikova et al., 1989) and purified it (Belozersky et al.,

    B.L. Jones / Journal of Cer1461990). The enzyme hydrolysed buckwheat storageproteins (Dunaevsky et al., 1983) and was inhibited by

    an endogenous inhibitor (Dunaevskii et al., 1995). Wrobel

    and Jones (1993) then showed that 4-day-germinated

    barley seed extracts contained five high Mrmetalloproteases

    5.4.1. Purification and characterization of malt

    metalloproteases (MPs)

    Chromatography and chromatofocusing were used by

    Fontanini and Jones (2001) to isolate a group of metallo-

    proteases from green malt. The metalloprotease mixture

    (MP) was separated by PAGE and 2-D IEF!PAGE ongelatin-containing gels and its individual components

    studied. All the MP enzymes were maximally active at

    pH 78 and were inhibited by the chelating agents EDTA

    and o-phenanthroline, but not by other class-specific

    proteinase inhibitors. The activities of EDTA-inhibited

    enzymes were restored by the addition of low concen-

    trations of either Co2C, Mn2C or Zn2C ions, but they were

    inhibited by higher concentrations of these same ions. In

    nature, they probably contain Zn2C ions at their active sites.

    The MP mixture was separated by 2-D electrophoresis into

    three major and six minor components, all of which behaved

    like metalloproteases.

    The MPs were located mainly in the aleurone tissues of

    the malt and were present in only very small amounts in

    ungerminated barley seeds and during the first day of

    malting. The MP hydrolysed the D component of a hordein

    preparation in vitro at pH 4 much faster than it did the B or C

    hordeins. A purified D hordein preparation and its C

    hordein contaminants were hydrolysed at pH 4, with the

    D hordein being converted into smaller, although still large,

    fragments. The C and D hordeins were degraded quickly at

    pH 8 (D, 40 min; C, 120 min) but no hydrolysis occurred in

    the presence of either EDTA or o-phenanthroline.

    The results obtained with the isolated metalloproteases

    differed from those gotten with crude extracts. Enari and

    Mikola (1968) reported that 31% of the proteinase activity

    of barley was due to metalloproteases, but that in malt

    they accounted for only 9% of the activity, whereas none

    of the MP enzymes studied by Fontanini and Jones (2001)

    were present in unmalted barley. Sundblom and Mikola

    (1972) reported that the metalloproteases were synthesized

    in, and secreted from, the barley aleurone, and Zhang and

    Jones (1995b) detected metalloproteases in the starchy

    endosperm of green malt using 2-D electrophoresis.

    However, no MP were detected in malt endosperm

    tissues. One explanation for these divergent results is

    that Mikolas group and Zhang and Jones may have

    extracted metalloproteases that differed rather strikingly

    from those present in the MP mixture. Alternatively, the

    results could be explained if the endosperm and seed

    contained endogenous metalloprotease inhibitors, as

    reported by Enari and Mikola (1968). In that case, even

    if the enzymes were present, they might not have been

    cience 42 (2005) 139156detected because they were complexed with the inhibitors.

  • indicated that they are similar to cucumisin (Kaneda and

    Tominaga, 1975), but they differed strikingly from each

    other in their pI values and temperature stabilities. SEP-1

    during the mashing phase of brewing (see Section 4). It is

    real SThat such metalloproteinase inhibitors do occur in grains

    and that they play an important part in storage protein

    hydrolysis was shown by the studies with buckwheat

    (Dunaevskii et al., 1995). The presence of metalloprotei-

    nase inhibitors in malt might be one reason why this

    group of enzymes has been so hard to extract and study in

    the past. It could also help explain why only relatively

    low metalloproteinase activities are detected in malt

    extracts when recent studies (Jones and Budde, 2005)

    indicated that these enzymes play a major role in the

    release of soluble protein during malting and mashing (see

    Section 7.5).

    5.5. Barley serine proteases

    Enari and Mikola (1968) found that specific serine

    protease inhibitors caused no inhibition of the endoproteo-

    lytic activity of green malt extracts and concluded that no

    serine class proteases were present. However, 2-D electro-

    phoretic separations and analyses of malt extracts revealed

    the presence of several serine proteases (Zhang and Jones,

    1995a) and showed that one of them was especially

    abundant after 2 days germination (Zhang and Jones,

    1995b).

    5.5.1. Hordolisin

    Terp et al. (2000) purified and characterized a serine

    endoproteinase from green malt that they called hordolisin.

    Its Mr of 74,000, and pI of 6.9, suggested that it was

    probably one of the B group proteases identified by Zhang

    and Jones (1995a). Its inhibition characteristics indicated

    that it was a serine class protease and its N-terminal amino

    acid sequence was similar to that of cucumisin, a subtilisin-

    like melon (Cucumis melo) enzyme. The enzyme had a pH

    optimum of 6 and was remarkably heat stable. A thorough

    study of its specificity using synthetic peptide substrates

    showed that its specificity was similar to that of savinase or

    subtilisin BPN. Incubation of hordolisin with barley protein

    bodies for 24 h released very little hordein material. No 24 h

    controls were shown, so it may be that even the apparent

    small loss of hordein that did occur was due to protein

    precipitation, as reported by Fontanini and Jones (2001),

    rather than hydrolysis. In any case, it appears that this serine

    proteinase does not play any appreciable role in solubilizing

    hordein proteins during malting.

    5.5.2. SEP-1

    Fontanini and Jones (2002) purified and studied what was

    apparently the major green malt serine proteinase detected

    earlier by Zhang and Jones (1995a). The purified enzyme,

    named serine endopeptidase-1 or SEP-1, was present in

    small quantities, but had a high specific activity for gelatin

    and was easily detected on zymograms. Of many protease

    inhibitors tested, including those specific for trypsin and

    chymotrypsin, only PMSF and APMSF affected the

    B.L. Jones / Journal of Ceenzyme, confirming that it was a serine class enzyme.apparent that most of these methods have major drawbacks.

    Two of the most important problems were that each of

    substrates used is susceptible to hydrolysis by only a few

    proteases and that the methods are unable to distinguish

    between the individual proteolytic components in extracts.

    To investigate any particular enzyme in an extract, it had to

    be purified from other contaminating proteases. Such

    purifications are time consuming and costly and whenever

    several similar enzymes are present, which is common in

    malt (see preceding sections), it is often impossible to obtain

    completely pure enzymes. Additionally, there is always a

    possibility that the proteases will be altered during theappears to be the A1 activity of Zhang and Jones (1995a),

    whereas hordolisin is probably one of the B activity group.

    If this is so, then SEP-1 should have been present in the

    endosperm of malt, even if it did not hydrolyse storage

    proteins. Since neither SEP-1 nor hordolisin hydrolysed

    hordein preparations and SEP-1 was unable to hydrolyse

    any of the common barley seed protein classes, it seems

    very unlikely that they solubilize storage proteins during

    germination. This agrees with the finding that none of the

    serine proteases solubilized proteins during mashing (Jones

    and Budde, 2005). SEP-1 and hordolisin probably have

    protein processing roles like those of other plant serine

    proteases. The fact that they increase strongly during seed

    germination implies that they do play an important role

    during the initiation of growth of new plants.

    6. Detecting proteolytic enzymes in barley/malt

    Many assays, using several substrates, have been used to

    detect endoproteolytic enzymes in barley and malts, andSEP-1 was active between pH 4 and 7, with maximal

    activity at pH 5.56.5, and between 50 and 60 8C. Theenzymes N-terminus was blocked, but sequencing of

    internal peptides indicated that its primary structure was

    similar to those of the cucumisin-like enzymes.

    Ungerminated seeds contained no SEP-1 activity or

    protein, but both were present after 2 days germination and

    persisted through at least 6 days germination. The enzyme

    was present only at very low levels in the scutellum/embryo

    of dormant seeds, but increased between 2 and 6 days of

    germination in all the tested tissues except the starchy

    endosperm, where it was never detected.

    5.5.3. Neither hordolisin nor SEP-1 is involved with protein

    solubilization during germination

    The characteristics of both hordolisin and SEP-1

    cience 42 (2005) 139156 147multiple-step purification processes.

  • eal S6.1. An in solution quantitative assay

    We set out (Jones et al., 1998) to define a system that

    could be used to study quantitatively individual proteases or

    their mixtures and to quickly and efficiently separate the

    individual proteases present and to measure their activities,

    in at least a semi-quantitative manner. The assay utilized the

    colored substrate protein, azogelatin, in an in solution

    assay for measuring the activities of samples that contained

    either individual enzymes or mixtures. The azogelatin

    substrate has several excellent attributes: (1) it is a protein

    and thus should give a realistic view of how the enzymes act

    on a natural substrate; (2) it is readily hydrolysed by many

    proteases, in contrast to several other substrate proteins that

    were digested by only a few; (3) it is hydrolysed by enzymes

    of all four of the common protease classes, although it is not

    as susceptible to attack by the aspartic proteases as it is to

    those of the other classes (Jones et al., 1998); (4) its

    hydrolysis products are colored red and absorb light at

    440 nm, so they can be monitored without derivatization.

    This also obviates problems that are associated with

    measuring the release of peptides at 280 nm, where

    contaminating amino acids, proteins, etc. also absorb. The

    hydrolyses are also more reproducible, since only the

    hydrolysis of the azogelatin substrate is detected, whereas

    the commonly used 280 nm wavelength also detects the

    hydrolysis of any contaminating proteins, each of which

    might be hydrolysed at a different rate; (5) it is readily

    soluble between pH 3 and 10.5, the range of the proteolytic

    activity pH optima of the four malt proteinase classes; (6) it

    is readily precipitated with trichloroacetic acid; and (7) the

    azogelatin derivative is easily and reliably pared from

    porcine skin type A gelatin of w300 bloom (SigmaChemical Company, Cat No. G2500), by carefully follow-

    ing the protocol in Jones et al. (1998).

    It should be noted that most of the enzyme reaction rates

    reported in Jones et al. (1998), involving several different

    proteinase types, were linear for 30 min or less, so that only

    reactions performed for 30 min or less would yield true

    initial reaction rates. The main drawback to the method is

    that even though azogelatin and gelatin are hydrolysed by

    pepsin, an aspartic protease, the hydrolyses proceed at a

    much slower rate than they did with proteases of the other

    three classes (Jones et al., 1998). This, together with the data

    gathered from enzymes separated by 2-D electrophoresis

    that were used to hydrolyse azogelatin (see Section 6.2)

    indicates that neither gelatin nor azogelatin are particularly

    good substrates for measuring the activities of many of the

    aspartic proteases. Nonetheless, the susceptibility of

    azogelatin to hydrolysis by so many of the malt proteases

    makes it the substrate of choice.

    It would have been preferable to use underivatized

    gelatin as substrate rather than its azo derivative but that was

    not practical, because gelatin is not precipitable with TCA

    and the reaction products must be measured at or near

    B.L. Jones / Journal of Cer148280 nm. Therefore, in gelatin hydrolyses controls must berun for every individual reaction that contains different

    amounts of protein.

    6.2. A two-dimensional IEF!PAGE analysis system

    In many situations where malt proteases are being

    assayed more than one active enzyme is present but

    measurements of the individual component enzymes, rather

    than the overall activity, are required. In this situation, the

    component proteases are best separated using a 2-D gel

    system, after which the separated fractions can be assessed.

    Based on the findings of Wrobel and Jones (1992) that

    active enzymes of germinating barley could be partially

    separated by non-denaturing PAGE in gels that contained

    substrate protein, such a separation system was developed

    (Zhang and Jones, 1995a). After the separation, the gels

    were developed by allowing the separated enzymes

    detected to hydrolyse the incorporated protein substrate

    and staining the gel for protein. The areas containing the

    separated activities were detected as clear areas, where the

    substrate protein had been digested, against a blue back-

    ground of stained protein. By separating the proteases on an

    IEF tube gel, incorporating the IEF gel into the top of the

    PAGE gel slab and then performing the gel separation and

    development, it was possible to detect many different malt

    proteolytic enzymes on a single gel (Zhang and Jones,

    1995a).

    When gelatin or azogelatin was incorporated into the

    PAGE gel the advantages of these substrates, as outlined in

    Section 6.1, could be exploited. Incubating the gel slabs in

    solutions of varying pH values and/or that contained class

    specific inhibitors allowed the pH optima of the separated

    enzymes and their hydrolytic classes to be readily

    determined (Zhang and Jones, 1995a). For analyses of the

    aspartic proteases, PAGE gels containing edestin were used.

    Using the gelatin and edestin systems together, over 40

    separate protease enzymes were detected in green malt

    (Zhang and Jones, 1995a). The 2-D method yielded very

    reproducible separation patterns in experiments performed

    at different times. Generally, the extent to which the

    incorporated substrate was cleared from the gel was

    proportional to the activity, and semi-quantitative data

    could be obtained. It was very easy to differentiate between

    very low, low, medium, strong and very strong activities.

    This resolution was, of course, lost if too much extract was

    loaded onto the gels, so that the substrate was completely

    hydrolysed from some areas.

    This method was used to determine when the various

    malt endoproteases which were active at pH 4.8 (the

    apparent pH of germinating barley seeds) appeared during

    malting and where they were located in malted kernels

    (Zhang and Jones, 1995b). The results corroborated and

    extended those of many earlier experiments, showing that

    ungerminated and steeped barleys contained few endopro-

    teases, but that many components appeared within 2 days of

    cience 42 (2005) 139156germination. The endosperm, where presumably most of

  • hydrolysate by PAGE. This has shown more promise, but

    the peptides released are too small to be detected on gels

    2003b). Remarkably, a number of the enzymes were, in fact,

    real S(Marchylo and Kruger, 1985; Poulle and Jones, 1988). This

    method has, however, been used by several researchers

    (Davy et al., 1998; Koehler and Ho, 1990a,b; Poulle and

    Jones, 1988) to show that their purified proteases can

    probably hydrolyse barley storage proteins in vivo.

    It would be advantageous to be able to incorporate a

    well-characterized hordein into IEF!PAGE 2-D gels, ashas been done with gelatin, so that the individual

    components of proteinase extracts could all be tested at

    once. This experiment has been attempted, but has never

    been performed successfully. There are, however, reports

    that 1-D PAGE gels containing incorporated hordeins have

    been used to partially separate and analyze enzyme extracts

    (Kaneda and Tominaga, 1975; Wrobel and Jones, 1992; Drthe results obtained are only semi-quantitative and most ofstorage proteins are hydrolysed, contained representatives

    of each of the serine-, aspartic- and metalloproteinase

    classes and numerous cysteine proteases. The formation and

    cellular locations of the various proteinase types and their

    significance for malting and mashing could thus be

    discerned.

    An important initial finding from the 2-D studies was that

    within each protease class the individual members generally

    had similar pH optima but that between classes there were

    major differences. The cysteine and aspartic proteases were

    most active at pH values between 3.8 and 4.5, whereas the

    serine- and metalloproteases were optimally active at pH

    levels from about 6.0 to 8.5 (Zhang and Jones, 1995a).

    6.3. Use of barley proteins as substrates

    6.3.1. Hordeins

    It would, of course, be highly desirable to use barley

    storage proteins as substrates in endoproteinase assays and,

    as reported above, that has times been done. However, in

    addition to having to physically alter the hordeins to extract

    them from the grain matrix, other problems also arise. An

    in solution assay that used, hordein preparations as

    substrates (Baxter, 1976) differentiated exopeptidase and

    endoproteases but was very complicated and susceptible to

    error. The assays had to be conducted at around pH 3 to keep

    the substrate in solutions. These conditions are very

    different from those that exist during malting, brewing or

    seed germination. The internal pH of germinating barley

    seeds is 4.8 and the pH of North American mashes is 6.0. Six

    different hordein preparations were used by Baxter, and

    their hydrolysis rates varied by nearly 30-fold. This raises

    questions about which, if any, of these results is most

    indicative of the hydrolysis of the natural hordein substrate

    and the reasons for the variable results.

    Another approach has been to hydrolyse hordeins in

    solution (or in suspension) and then analyze the resulting

    B.L. Jones / Journal of CeMark Schmitt, personal communication). In these 1-D gels,activated by various inhibitors, by up to 60%. Experiments

    with what was apparently the same substrate, now called

    glutelin, indicated that about 60% of the endoproteolytic

    activity of malt was present after one day of germination,

    whereas analyses using gelatin-containing gels or the

    hydrolysis of haemoglobin both indicated that only about

    10% of the activity was present at that time (Osman et al.,

    2002). Compared to the seven characteristics of gelatin

    listed in Section 6.1, this substrate does not meet criterion 4,

    has not been tested for criteria 2, 3, and 5, and its

    compliance with criterion 7 is questionable. It is unlikely

    that this substrate will prove useful for protease assays.

    7. Understanding in vivo proteolysis

    As seen from Section 5, we now have a reasonable

    knowledge of the characteristics of several of the barley/

    malt proteases, but we still need a better understanding of

    what really happens in barley at pH 4.8 and in mashes at pH

    6.0. This information is needed by researchers so that they

    can rationally alter the barley protein hydrolysis system to

    produce improved malting or germinating barleys and by

    maltsters/brewers so that they can adjust their processing

    methods to prepare improved malts and beers.

    7.1. The formation of proteolytic enzymes during malting

    and their stabilities to kilning

    Since the earliest studies, it has been clear that the overall

    endoproteolytic activity of barley grains is quite low and

    that proteases were formed and/or activated during the early

    phases of seed germination or malting (Harris, 1962). The

    real question was, therefore, which proteases were formedthe hordein substrate was not hydrolysed nearly as readily as

    gelatin.

    6.3.2. Highly degradable barley protein fraction (HDBPF)

    Osman (2003a) proposed the use of a barley preparation

    known as highly degradable barley protein fraction

    (HDBPF), as a natural substrate for malt endoproteinase

    assays. There are several problems with this substrate,

    however. Its hydrolysis rate was not linearly correlated with

    the amount of added enzyme and the reaction rates actually

    dropped as the substrate levels were increased. In addition,

    the HDBPF has none of the characteristics of normal

    hordein storage proteins (Mr too low, amino acid compo-

    sition completely different, etc.). In five malt proteinase

    mixtures assayed using this substrate with class specific

    protease inhibitors, the results suggested that none of the

    enzymes present were either cysteine- or metalloproteases.

    All were completely inactivated by the serine class

    inhibitor, DIC, and two were also inhibited by 98% by

    pepstatin A and may have been aspartic proteases (Osman,

    cience 42 (2005) 139156 149and what did they do? These questions could only be studied

  • teinase activities remained constant throughout the 30 min

    eal Safter methods had been developed for quickly and efficiently

    separating the various enzymes and measuring their

    activities under conditions similar to those that occur in

    the germinating seed and in mashes. By quantifying the total

    endoproteolytic activities of samples using the azogelatin

    in solution assay (Jones et al., 1998) and semi-quantitat-

    ively analyzing the component enzymes with the 2-D

    method, using azogelatin or gelatin as substrate, it became

    possible to see how changes in individual enzymes affected

    the overall protease activity.

    These methods were first applied to the study of barley

    undergoing malting (Jones et al., 2000). As expected, under

    malting conditions the proteolytic activity at the end of

    steeping was very low, but began to rise at germination day

    1 and was maximal by day 3. The activities at pH 3.8

    (measured for comparison with the results of many earlier

    studies), at pH 4.8 (the internal pH of inside germinating

    grain) and at pH 6.0 (mashing pH) increased concomitantly.

    The assays at pH 3.8 would have measured only the

    cysteine- and aspartic proteases and the assays at pH 6.0

    mainly the serine- and metalloproteases (Zhang and Jones,

    1995a, see Section 6.2), so the enzymes of the different

    classes must have increased at the same time.

    Sampling of green malt during the kilning phase of

    malting showed that there was no diminution of proteolytic

    activity during kilning at any of the three measured pH

    values, even when the temperature was raised to 85 8C. Inthe absence of added cysteine, the kilned malt activities

    were even slightly higher than those of green malt (Jones et

    al., 2000). 1-D PAGE analyses confirmed these findings and

    showed that the same proteases were active throughout the

    kilning process, but that completely different sets of

    enzymes were active at pH 3.8 and 6.0. Analyses using

    the more sensitive 2-D system, however, showed that some

    minor changes did occur in some cysteine and serine

    proteases during the 68 and 85 8C steps. Apparently, therewas a partial denaturation of the enzymes that resulted in the

    enzyme spots on the gel becoming more diffuse. It was

    obvious that there was no loss of proteinase activity during

    kilning, even upon heating to 85 8C.

    7.2. The effect of mashing on proteolytic activities.

    Mashing is the first step in the brewing process, when

    milled malt is subjected to a carefully controlled extraction

    with water whose temperature is gradually raised. The effect

    of mashing on the malt proteinase activities was tested using

    a mash regime that was based on US commercial methods

    (Jones and Marinac, 2002). In solution assays showed that

    the overall proteolytic activity was constant throughout a

    50 min, 38 8C protein rest phase, but fell rapidly when thetemperature was raised to 72 8C for the conversion phase.The results were the same at pH 4.8 and 6.0, so the

    components of all of the enzyme classes behaved similarly.

    These results were confirmed by 1-D PAGE analysis, which

    B.L. Jones / Journal of Cer150also confirmed that quite different enzymes were active atprotein rest, but the overall activity dropped more than

    5-fold as the pH was raised from 5.1 to 6.6. This reflects the

    change from cysteine proteases to serine- and metallopro-

    teases as the predominant active enzymes. The wort soluble

    protein and FAN levels changed in concert with the

    proteolytic activities, in a sigmoidal manner. In addition,

    the extract values dropped by 4% points which, for a

    commercial malt, is a very large change. The (1/3,1/4)-b-glucan levels increased as the pH was raised, butremained at commercially acceptable levels.

    It was thus possible to produce worts with widely varying

    compositions simply by changing the mashing pH. It would

    be interesting and instructive to see how using worts with

    these non-traditional FAN, SP and extract contents would

    affect the brewing process. The differences in the proteolytic

    activities at pH 4.8 and 6.0 indicate that during malting and

    mashing the pattern of hydrolysis of the storage proteins

    must be very different, and that measurements made in onepH 4.8 and 6.0. 2-D PAGE analyses verified the 1-D results

    and defined which protease types were active at each pH.

    All four enzyme classes were active at pH 4.8, but

    metalloproteases predominated at pH 6.0, although the

    major serine enzyme was also active. During mashing, all of

    the enzymes were inactivated simultaneously and the

    inactivation of the individual enzymes correlated well

    with the loss of overall activity, as measured in solution.

    That the enzymes have practically identical heat labilities

    implies that it will not be possible to alter the amino acid

    compositions of mashes by changing the mash temperature

    regime. However, the soluble protein levels of worts can

    obviously be increased by extending either the malting

    process or the protein rest phase of mashing, but not by

    extending the mash conversion time. The great majority of

    green malt proteases were much more heat stable in the malt

    kernel (stable to 85 8C for over 3 h during kilning) than in

    solution (inactivated within 5 min at 72 8C).

    7.3. The effect of pH on malt and mash proteases

    and on worts

    Because the pH optima of the different malt protease

    enzyme classes are so different, it should be possible to vary

    the compositions of worts by changing the mashing pH.

    This was tested by conducting room temperature mashes

    at initial pH values from 4.4 to 7.1 (Jones and Budde, 2003).

    The mashes had good buffering capacities and their pH

    values adjusted quickly towards pH 5.8, the apparent natural

    mash pH value, so that the final mash pH values ranged from

    4.8 to 6.4.

    The wort characteristics of mashes performed at 45 8C

    and between pH 5.1 and 6.6 were very different (Jones and

    Budde, 2003). At each pH value, the individual endopro-

    cience 42 (2005) 139156system cannot be presumed to apply to the other.

  • real S7.4. The effects of reducing and oxidizing agents

    on proteases in malts, mashes and worts

    Reducing agents (most often 2-ME) have routinely been

    added to nearly all barley and malt proteinase preparations

    for decades because they allegedly increased the amounts of

    extractable enzymes and/or maintained their activities.

    Some of this reducing agent was usually transferred into

    assay mixtures along with the enzyme and additional

    reducing agent was often added directly to the proteolytic

    analyses to maximize the activities. These practices raise the

    question of whether the results thus obtained provide a true

    indication of what really happens during mashing and

    brewing.

    Apart from the effect of exogenous reducing agents,

    Kobrehel et al. (1991, 1999) have proposed that endogenous

    oxidation/reduction systems may control the reduction

    states of various grain proteins and thereby affect the

    operation of some of the grain proteases. The evidence for

    this is circumstantial, but it needs to be investigated.

    The effects of adding reducing and oxidizing agents to

    malt extracts have recently been studied (Jones, 1999; Jones

    and Budde, 2003). When considering both the studies of

    Jones and Budde and those of Buchanan and his

    collaborators (Kobrehel et al., 1991, 1999) it must be

    remembered that redox reagents, whether exogenous or

    endogenous, may have two distinct effects. They may

    activate/inactivate cysteine proteases and/or they may alter

    the substrate molecules, making them more or less

    susceptible to hydrolysis.

    The addition of cysteine to mashes increased their

    proteolytic activities by 3- to 4-fold (Jones, 1999; Jones and

    Budde, 2003). Other amino acids had no effect, indicating

    that the cysteine effect was due to its reducing power. The

    addition of weak (diamine) or strong (hydrogen peroxide,

    H2O2) oxidizing agents to malt extracts lowered their

    proteolytic activities and when diamine or H2O2 was added

    to reactions together with cysteine, the cysteine effect was

    cancelled, confirming that the effects were due to redox

    reactions. With proteinase assays in solution, the addition of

    low concentrations (up to 1 mM) of the strong reducing

    agent dithiothreitol (DTT) led to an enzyme activation that

    disappeared as the DTT concentration was increased

    further. On the other hand, the proteolytic activity was not

    affected by the addition of 2-ME.

    Conversely, for individual proteases that were separated

    and analyzed using the 2-D system cysteine, DTT and 2-ME

    all strongly activated certain cysteine class proteases, and all

    of these activations were negated by diamine (Jones and

    Budde, 2003). No serine- or metalloproteases were affected

    by any of the redox agents. The addition of cysteine, DTT or

    2-ME to ASBC mashes led to increases in their wort SP,

    FAN and extract values. The presence of either diamine or

    H2O2 in mashes produced some reduction in their SP and

    FAN levels, and both oxidizing agents effectively

    B.L. Jones / Journal of Cenegated the increases caused by the three reducing agents.The extract (1/3,1/4)-b-glucan levels were unaffectedby any of the redox agents.

    7.5. Which proteases actually release SP and FAN

    during malting and mashing?

    To better define which endoproteases effect changes in

    the SP and FAN levels of brewhouse worts, experiments

    were performed in which class-specific proteinase inhibitors

    were added to Morex (6-rowed) and Harrington (2-rowed)

    barley malt mashes (Jones and Budde, 2005). At pH 6.0,

    the efficacy of inhibitors in lowering the wort SP levels

    was o-phenanthrolineOE-64Opepstatin AOPMSFZ0,indicating, surprisingly, that the metalloproteases were

    responsible for controlling the solubilization of more

    protein during mashing than were the cysteine enzymes,

    and that the aspartic proteases also played a significant, if

    lesser, role. These results were confirmed with mashes

    conducted at pH 3.8 (cysteine and aspartic proteases active)

    and pH 8.0 (serine and metalloproteases active), even

    though the control SP levels were greatly increased at pH

    3.8 and reduced at pH 8.0. These metalloproteases results

    were based on inhibition with o-phenanthroline, since

    EDTA strongly disrupted the system and, in most cases,

    led to increased, not lowered, protease activities, SP levels

    and FAN contents. The SP, extract and FAN levels were all

    strongly affected by pH, but proteinase inhibitors did not

    affect the wort extract values. The presence of inhibitors

    generally lowered FAN levels at pH 3.8 and 6.0, but the

    effect was not dramatic. The inhibitory effects did not vary

    with either malt concentrations or mashing conditions. The

    SP, extract and (1/3,1/4)-b-glucan levels varied amongthe different pH treatments, but inhibitors had no effect on

    wort extract or (1/3,1/4)-b-glucan levels and theinhibition of the cysteine and metalloproteases most

    strongly lowered the wort SP concentrations.

    These experiments led to the conclusion that malt

    metalloproteases play a much more prominent role in the

    solubilization of protein during mashing (and possibly also

    during malting) than was previously suspected. It also

    appears that, notwithstanding the characteristics of those

    aspartic proteases purified and characterized to date, these

    proteases also play a significant role during mashing. In rye,

    the aspartic proteases apparently play a major role in

    hydrolyzing storage proteins during seed germination (Brijs

    et al., 2002), so it is not too surprising that they are also

    important in barley germination.

    7.6. When is protein solubilized during seed germination?

    Several attempts have been made to determine the

    percentage of the wort SP that is released during the separate

    malting and mashing procedures. For mashing, values

    reported have varied from 0% (Lewis et al., 1992) to 30%

    (Burger and Schroeder, 1976a) and 47% (Barrett and Kirsop,

    cience 42 (2005) 139156 1511971). By mashing ungerminated barley and malt in

  • malting and mashing (Jones et al., 2000; Jones and Marinac,

    7.8.

    of v

    V

    prot

    metalloproteinase (Dunaevsky et al., 1983), after which a

    and Jones, 2000a), even though cysteine proteases appar-

    ently degraded most of the storage protein (Mikola and

    Jones, 2000b). In germinating rye seeds, the protein

    eal Science 42 (2005) 1391562002). The proteolytic activities of six Australian and North

    American barleys, tested with three different substrates, also

    did not differ greatly (Osman et al., 1997, 2002). On the

    other hand, analyses of malts made from 43 barley lines

    from around the world showed that their proteinase values

    varied by 4- to 6-fold, that the cysteine and aspartic protease

    activities predominated, and that only the cysteine protein-

    ase activities correlated with their SP values or Kolbach

    Indexes (Kihara et al., 2002). For a number of reasons these

    latter results are questionable; the pH values of the protease

    assays were not specified, 5 mM DTT was added to the

    assays, which would have greatly increased both the

    cysteine class proteinase activities and the wort soluble

    protein levels (Jones and Budde, 2003), and a casein

    derivative was used as substrate, even though casein is not

    hydrolysed by most malt endoproteases (Jones et al., 1998).

    Moreover, initial reaction rates were not used for the kinetic

    analyses (see Section 4). From the high aspartic- and low

    metalloprotease levels that were reported, it seems likely

    that the assays were conducted at pH 5.0 (Zhang and Jones,

    1995a), which would have strongly enhanced their overallsolutions that contained mixtures of protease inhibitors,

    Jones and Budde (2005) measured the SP levels of the barley,

    malt and wort, and calculated that, at pH 6.0, 32% of the final

    wort SP was already present in unmalted barley, 46% was

    solubilized during malting and 22% during mashing. At pH

    8.0, where the cysteine and aspartic proteases were inactive,

    10% of the total wort SP was solubilized during mashing,

    another indication of the major contribution of the metallo-

    proteases to the SP levels. Because this study measured, for

    the first time, the SP levels of the unmalted grain and used

    relatively benign specific protease inhibitors, the results

    should be particularly meaningful. The results obtained

    would presumably vary if different malting and mashing

    regimes were used.

    Of the FAN in wort, 15% was present in unmalted barley,

    58% was released during malting and 26% during mashing

    (Jones and Budde, 2005). The difference between the FAN

    and SP results was expected since the FAN and SP are

    released by different sets of enzymes. At pH 8.0, no FAN

    was released during mashing, implying that none of the malt

    exopeptidases were active at this high pH.

    7.7. Variation among barley cultivars

    Jones and collaborators used both 6-rowed (Morex) and

    2-rowed (Harrington) barleys in their proteinase exper-

    iments. Both are very good malting cultivars although they

    have significantly different characteristics. Their proteolytic

    profiles were the same and they responded identically to

    redox compounds and changes in pH (Jones and Budde,

    2003), to inhibitors (Jones and Budde, 2005) and during

    B.L. Jones / Journal of Cer152proteolytic activities and FAN and SP levels.breakdown was mostly catalyzed by the cysteine and

    aspartic enzyme classes (Brijs et al., 2002).

    A comparison of the proteolytic complements in malted

    grains of barley, bread and durum wheats, rye, triticale, oats,

    rice, buckwheat and sorghums was made by separating their

    proteases and analyzing them using the gelatin 2-D system

    (Jones and Lookhart, 2005). Their IEF and PAGE migration

    characteristics and the effect of pH changes allowed an

    estimate to be made of the members of the various enzyme

    classes that were present in each species. All of the

    germinated grains contained multiple enzymes. The separ-

    ation patterns and pH characteristics of the bread and durum

    wheats, ryes, oats and sorghums were fairly similar to those

    of barley, whereas the patterns in other grains showed more

    variability. Rice and buckwheat proteases developed very

    slowly. In triticale the activity patterns were similar to those

    of their wheat and rye parents, but the triticales contained

    many more proteases and their overall activities were the

    highest of any of the species that were tested.

    These results complement previous findings that indi-

    cated that cereal grains tend to contain similar proteases, but

    that each species may degrade its storage proteins

    differently. One important finding was that all of the cereals

    except rice exhibited strong metalloproteinase activities,

    supporting the proposal that these enzymes play a greater

    role in the degradation of grain proteins than has been

    previously assumed.

    8. The current situation

    In overviewing the current state of proteinase studies in

    barley and malts, several things stand out. Among these are:

    (1) Cysteine proteases are clearly important players in the

    hydrolysis of barley proteins during malting and

    mashing. However, it seems likely that they do not

    play as predominant a role as was attributed to them in

    the past. Most earlier studies were performed at pH

    values below 4.8 and in the presence of added reducingcysteine proteinase degradation predominates (Dunaevsky

    and Belozersky, 1989), possibly with the assistance of an

    exopeptidase. In oats, most of the at proteases pH 6.2 active

    present initially were serine and metalloproteases (MikolabucVariations in the proteolytic capacities

    arious grain species

    arious grain species apparently hydrolyse their storage

    eins differently during germination. For example, in

    kwheat seeds the initial hydrolysis is catalyzed by aagents, conditions that strongly increase the cysteine

  • (2)

    (5)

    Dividing the multiple malt proteases into five groups

    of enzymes without showing that these fractions contain

    I

    res

    that

    Thi

    of m

    199

    can

    con

    hav

    stor

    Unf

    rese

    I

    hav

    stim

    my

    Ms

    Wro

    B.L. Jones / Journal of Cereal SAny new studies of barley/malt proteases should

    concentrate on the aspartic and metalloproteinasebeen purified and characterized seem to be involved in

    hydrolysing the seed storage proteins, it is likely that

    other members of this group do participate. Four

    aspartic proteases were detected in green malt by 2-D

    separations (Zhang and Jones, 1995a) and Zhang and

    Jones (1999) showed that at least six forms occurred in

    ungerminated seeds, but none of the seed enzymes

    hydrolysed barley hordein prepara