Heterologous expression and protein engineering of wheat gluten proteins

Journal of Cereal Science 43 (2006) 259–274 www.elsevier.com/locate/jnlabr/yjcrs

Heterologous expression and protein engineering of wheat gluten proteins Laszlo Tama´s a, Peter R. Shewry b,* a

Eotvos Lorand University of Sciences, 1/C Pazmany Peter Stny, Budapest H-1117, Hungary b Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK

Received 24 November 2005; received in revised form 8 February 2006; accepted 11 February 2006

Abstract A range of systems are available for the production of recombinant wheat gluten proteins, from simple and widely used systems based on Escherichia coli to more sophisticated eukaryotic systems in yeasts or cultured insect cells. The characteristics of these systems are summarised and their advantages and disadvantages for application to wheat gluten proteins discussed. We then review the applications of heterologous expression systems to the synthesis and characterisation of wheat gluten proteins, including the production of wild type and mutant proteins for structure–function studies. We also discuss the use of heterologous expression to establish model systems including perfect repeat peptides based on motifs present in gliadins and glutenin subunits and ‘analogue glutenin proteins’ based on C hordein of barley. It is concluded that the pET series of vectors and E. coli are suitable for most applications, providing high-level expression and being rapid and easy to use. q 2006 Elsevier Ltd. All rights reserved. Keywords: Wheat gluten proteins; Protein engineering; E. coli; Recombinant proteins

1. Introduction The heterologous expression of proteins and protein engineering are often considered to be almost synonymous but in fact they are quite different in concept. Heterologous expression refers to the expression of a protein in an organism different to that from which it originates. In practical terms, this is usually a microorganism or cell type which can be readily grown in culture, with high cell densities facilitating the production of high yields of recombinant (i.e. expressed) proteins. However, expression systems are not limited to cells or organisms that can be cultured and in some cases it may be advantageous to use whole eukaryotic organisms such as plants as hosts. The most widely used microorganism is Escherichia coli but yeasts (Saccharomyces cerevisiae, Pichia pastoris) and filamentous fungi are also used. The characteristics of these systems and their relative advantages and disadvantages for the expression of gluten proteins will be discussed in detail in Section 2. Abbreviations: ANG, analogue glutenin proteins; CD, circular dichroism; ES-MS, electrospray mass spectrometry; FT-IR spectroscopy, Fourier-transform infra-red spectroscopy; MALDI-tof MS, matrix assisted laser description ionisation-time of flight; Q-tof MS, quadruple-time of flight mass spectrometry. * Corresponding author. Tel.: C44 1275 549348; fax: C44 1275 394299. E-mail address: [email protected] (P.R. Shewry). 0733-5210/$ - see front matter q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2006.02.001

Perhaps, the widest use of heterologous expression systems is to allow the production of single homogeneous proteins. In some cases, the protein encoded by a gene or cDNA has not been characterised, or even identified, at the protein level. A heterologously expressed protein after purification can be used to assay for biological activity in vitro or to raise antibodies that can, in turn, be used to determine its location and amount in vivo. In the case of cereal prolamins, the identification of individual proteins may not pose a problem but the purification of a single component from a highly complex mixture of related proteins frequently does. Hence, heterologous expression is most frequently used to produce single pure components for use in structure–function studies. Protein engineering is used to define structure–function (or structure–functionality) relationships by creating and characterising mutant forms of a protein. In the case of peptides and small proteins, the mutant forms can be created chemically by peptide synthesis. However, this becomes increasingly difficult and expensive as the length of the polypeptide chain increases and heterologous expression is the preferred route, with mutations being introduced into the DNA encoding the protein using standard genetic engineering technology. The wild type and mutant proteins are then expressed in an appropriate system, purified and characterised. Analysis of the recombinant proteins can provide information on the molecular basis for the functional properties of cereal grain proteins, which may have direct applications in developing optimal conditions for grain processing. However,

260

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

in many cases, it will be necessary to effect improvements via plant breeding, including wide crossing to introduce novel proteins from related species, or by genetic engineering. These approaches are expensive and time consuming and breeders are therefore reluctant to embark on new projects unless there is a sound scientific basis and reasonable chance of success. In such cases, heterologous expression and protein engineering can be used to allow the functional properties of individual proteins to be determined as a prelude to their selection or introgression by plant breeding, or introduction using genetic engineering technology. This review focuses on the application of heterologous expression systems to study the structures, interactions and functional and biological properties of wheat gluten proteins. However, before discussing this work in detail we will briefly describe the characteristics of the most widely used expression systems, focusing on specific factors which determine their appropriateness for expressing gluten proteins.

2. Characteristics of expression systems Three types of expression systems are widely used for heterologous protein production, exploiting E. coli, yeasts and cultured insect cells as hosts. Their characteristics and advantages/disadvantages are summarised in Table 1 and discussed briefly below. The Gram-negative bacterium E. coli is the most widely used microorganism for heterologous protein expression and many workers only consider other systems if yields from E. coli are inadequate or if the protein produced is biologically inactive. Expression systems developed from E. coli benefit from its well-characterised genetics and a large number of vectors and host strains are now available commercially allowing the precise system to be optimised for individual proteins. The copy numbers of E. coli expression vectors depend on their plasmid of origin with most vectors being derived from low copy number plasmids. For example, the widely used pET vectors (Studier and Moffatt, 1986) are based on the pBR322 plasmid which is present at 15–60 copies/cell. However, high copy number plasmids such as pUC (O200 copies/cell) have been used to construct vectors, such as pLEX (Mieschendahl et al., 1986), for high-level protein expression (see excellent reviews by Jonasson et al., 2002; Sørensen and Mortensen, 2005). Unicellular eukaryotic yeasts have been used for large-scale production of recombinant proteins since the early 1980s. Our detailed knowledge of yeast genetics and physiology makes these organisms particularly suitable for development as a eukaryotic expression system. In particular, yeast cells recognise eukaryotic translation, processing and modification signals and so perform eukaryotic processing steps on the polypeptides expressed. Most early studies used bakers’ yeast (S. cerevisiae) (reviewed by Hinnen et al., 1994) but an increasing number of alternative systems have been developed to avoid specific limitations of this species. They include the

methylotrophic yeast species Hansenula polymorpha, P. pastoris and Candida boidinii (Faber et al., 1995; Gellissen, 2000; Gellissen and Melber, 1996; Gellissen et al., 1992) and the budding yeast Kluyveromyces lactis (Bergkamp et al., 1992). The methylotrophic yeast species share a highly inducible methanol utilisation pathway and can therefore grow on methanol as a sole carbon and energy source. P. pastoris has proved to be an excellent host for protein production and has become increasingly popular due to a wide choice of host strains and vectors and its ability to be cultured in high density fermenters (Cereghino and Cregg, 2000; Daly and Hearn, 2005). However, so far Pichia has not been used for expression of wheat gluten proteins. Most yeast expression vectors can either exist autonomously in the cell or become integrated into the bacterial genome (and hence are often termed episomes rather than plasmids). They may be present either in low copy number (1–2) (such as YRp and YCp), or in high copy number (YEp), with the latter (maintaining about 30–100 copies per cell) being based on the 2 mm plasmid. Protein expression in insect cells was first reported in the early 1980s (Smith et al., 1983) using baculovirus as an expression vector. Baculoviruses constitute one of the most diverse groups of arthropod viruses with the best studied member of the family being Autographa californica, a nuclear polyhedrosis virus (AcMNPV) with a double stranded, circular DNA genome. The most commonly used insect host cell lines were originally derived from pupal ovarian tissue of Spodoptera frugiperda (the fall armyworm, a tropical lepidopteran species). The baculovirus system has been widely used and has signaficant advantages including the capacity to accept large inserts of DNA and the production of high yields of recombinant protein (for recent reviews see Ikonomou et al., 2003; Kost and Condreay, 1999; Philipps et al., 2005; Possee, 1997). However, the large size of the AcMNPV genome means that it is not possible to construct the viral expression vector using standard recombinant DNA technology with restriction enzymes. Instead, the foreign gene is inserted into a transfer vector flanked by viral sequences and used to coinfect cultured insect cells together with wild type AcMNPV. Recombination between the transfer vector and wild type virus occurs leading to the production of recombinant virus, which can be isolated and used to infect fresh insect cells for protein production. Thus, the production of the recombinant virus is a more time consuming and technically demanding process than the construction of vectors for E. coli and yeast expression. Furthermore, the multiplication of the recombinant AcMNPV, the determination of the virus titre and the optimisation of conditions are all time consuming. Nevertheless, once established the baculovirus system can give spectacularly high yields of recombinant proteins. Despite these clear advantages, there has only been one report of the expression of a wheat gluten protein using a baculovirus system (Thompson et al., 1994).

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

261

Table 1 Summary of the advantages and disadvantages of heterologous expression systems for wheat gluten proteins

Advantages

Escherichia coli

Saccharomyces cerevisiae (bakers’ yeast)

Pichia pastoris (yeast)

Baculovirus

Cheap, quick and easy to use

Glycosylation and proteolytic processing may occur Possess the endoplasmic reticulum (ER) machinery for protein folding and processing Protein folding may be authentic

Easy to manipulate, good secretory capacity High yield

High yield

Various strains and vectors are available Stable integration of vectors at specific sites into the genome in single or multicopy Fusion technology is available Simple to scale-up from shaking flask culture to high density fermenter Low risk of contamination by most other microorganisms Glycosylation and proteolytic processing may occur Possess the endoplasmic reticulum (ER) machinery for protein folding and processing Protein folding may be authentic

Glycosylation and proteolytic processing may occur Possess the endoplasmic reticulum (ER) machinery for protein folding and processing Protein folding may be authentic

Biased codon usage Promoters are not completely down regulated under non-induced conditions

Biased codon usage Slow, low yield

Biased codon usage Slow, expensive and requires specialised equipment

Protein often unfolds and precipitates in inclusion bodies S–S bond formation occurs only in specific host strains

Relatively low secretory capacity

Biased codon usage Accumulation of methanol may negatively affect cell growth and lead to decreased protein expression level Optimal methanol concentration needs to be determined for each expressed protein Glycosylation and proteolytic processing may not be authentic

Wide range of vectors and many different host strains are available High yield Fusion technology is available

Disadvantages

Glycosylation and proteolytic processing may not be authentic

Fusion technology is available

Glycosylation and proteolytic processing may not be authentic

No post-translational modifications Specific relevance to gluten protein expression

Inclusion bodies may facilitate purification

No glycosylation or proteolytic processing

Folding of cysteine-containing proteins may be more authentic with a proportion of the recombinant protein containing correct disulphide bonds There is no evidence of glycosylation of gluten proteins

High yields of highly repetitive proteins and peptides

3. Advantages and disadvantages of expression systems for application to gluten proteins Although many scientists routinely use E. coli systems, it is nevertheless wise to consider the advantages and disadvantages of the various hosts, vectors and promoters before attempting the expression of a new protein, particularly if the native protein contains intra-chain disulphide bonds and has unusual characteristics such as water-insolubility and repetitive structure.

The main considerations relate to protein yield, the authenticity of the expressed protein and the ease of purification and these are discussed separately below.

3.1. Protein yield Protein yield is clearly a major parameter to consider when establishing an expression system and is affected by a number of factors. Three of these are particularly relevant to the

262

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

expression of gluten proteins: choice of promoter, plasmid stability and codon usage. Expression vectors require a strong promoter for high-level protein expression. Although constitutive promoters are available for all systems, the most widely used promoters are inducible, with expression being induced by either specific chemicals or thermal effects. Inducible systems allow cultures to grow to high density before expression is induced, providing higher yield and minimising effects of toxic proteins. Nevertheless, some basal expression may occur in the absence of the inducer and this can be enough to result in cell death if the product is toxic. If so, tighter regulation can be achieved in bacteria by the addition of the lac operator sequence next to the promoter, and/or by the use of an appropriate repressor molecule. For example, expression of the lacI repressor protein down-regulates the production of the target protein, which only occurs after the expression of the lacI repressor is blocked (Mieschendahl et al., 1986). Tight regulation and the use of high-density cultures may therefore provide significant quantities of toxic protein even if the host cells ultimately die. A wide range of bacterial systems are available, using different inducible promoters. The T7-based promoter induced by IPTG (isopropyl-b-D-thiogalactopyranoside) is particularly popular for high-level expression. Other promoters induced by heat shock (lac(TS)) (Hasan and Szybalski, 1995), lowering of the temperature (cspA) (Vasina and Baneyx, 1997) and osmotic pressure (proU) (Bhandari and Gowrishankar, 1997) are also available. Strong promoters can also be used for expression of foreign protein in yeast systems. These promoters are derived from the alcohol oxidase 1 (AOX1) and glyceraldehyde-3phosphate dehydrogenase (GAP) genes. The promoter of the alcohol oxidase gene (AOX1) of P. pastoris is one of the strongest promoters known. It is tightly repressed by glucose and most other carbon sources but induced by over O1000fold when cells are transferred to methanol as a sole carbon source. Dai et al. (2005) have also described a baculovirus expression system based on the ecdyson receptor. The vector consists of chimaeric regulatory elements giving tight transcriptional regulation of foreign genes in insect cells. Plasmid instability can arise when the product of the gene is toxic to the host cell, when the plasmid is lost during cell division or when the gene is modified by recombination. Low copy number plasmids tend to be preferred if protein toxicity is a problem as the level of accumulation will be lower (but yields will also be reduced). If the toxic protein is accumulated in inclusion bodies, the issue of yield and plasmid instability can easily be solved, however, a new problem, protein authenticity, may arise (see below). Improved stability of the pET vectors which are widely used in E. coli has been achieved by expressing T7 lysozyme, a natural inhibitor of RNA polymerase, to reduce the level of basal expression which occurs without induction (Moffatt and Studier, 1987; Studier et al., 1990). Finally, harvesting of the cells only a short time after induction (2–3 h for E. coli) minimises toxic effects on cell growth and expression. Plasmid instability due to recombination is a greater risk for highly repetitive genes such as those encoding wheat gluten

proteins than for genes comprising only non-repetitive sequences. In some systems, recombination can be minimised by using host strains, which have mutations in their recombinase systems. Such recombinase-deficient hosts are available in E. coli, yeast and Pichia but not in the baculovirus system where recombination may give problems when repetitive sequences are expressed (as discussed below). Furthermore, recombination may still occur even if recombinase-deficient host strains of E. coli are used and it is wise to check plasmid authenticity between each round of expression. Amino acids are each encoded by between one and six different codons, and where multiple codons exist, their frequency of use may vary between different groups of organisms. If such differences in codon usage occur between the gene of interest and the host cell, the expression of the recombinant protein may be limited by the low abundance of the corresponding tRNAs. This could be a problem with plant proteins expressed in E. coli and particularly with gluten proteins because they contain high proportions of a small number of amino acids (notably proline and glutamine). However, the high yields achieved when a number of unmodified wheat gluten protein genes have been expressed in E. coli (see below) suggests that codon usage is not a limiting factor in practice. It would clearly be a formidable task to change the codon usage of a highly repetitive gluten protein gene. Nevertheless, codon usage has been optimised when constructing novel genes encoding gluten-related peptides. Thus, both Anderson et al. (1996b) and Elmorjani et al. (1997) optimised the codon usage to match that of the E. coli host cells while designing sequences encoding perfect repeat peptides corresponding to those present in the HMW subunits and gliadins, respectively. Similarly, Feeney et al. (2001) biased the codon preference to that favoured by E. coli when designing constructs to express repetitive peptides based on HMW subunits but also used alternative codons in some positions to minimise the repetition and hence maximise stability of the constructs. 3.2. Protein authenticity Heterologously expressed proteins must be correctly folded and processed if measurements of their biological activity and/or functional properties are to be meaningful. In many cases, folding includes the correct formation of intra-chain disulphide bonds, and this is required for most gluten proteins. Hence, protein authenticity is an important factor to consider when selecting hosts and vectors. High-level expression of disulphide bond stabilised proteins in E. coli almost inevitably results in the precipitation of the recombinant protein within the cell to form inclusion bodies (van den Berg et al., 1999). This appears to result from the absence of the required enzymes and chaperones that assist protein folding and disulphide bond formation in eukaryotic cells together with the reductive conditions in the cytoplasm of the bacterium, which do not favour disulphide bond formation. However, the rate at which inclusion body formation occurs can be modulated by controlling various expression

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

parameters, such as the temperature and rate of expression (Jonasson et al., 2002). Specific systems have also been developed to assist disulphide bond formation, including the use of host strains which are mutant in genes for cysteine reduction (trxB/gor) (Aslund and Beckwith, 1999; Zapun et al., 1995), expression of the recombinant protein as a fusion with thioredoxin (LaVallie et al., 1992) and secretion of the recombinant protein into the periplasm which is oxidative and contains enzymes of the Dsb system which catalyse disulphide bond formation and rearrangement (Rietsch and Beckwith, 1998). Co-expression of mammalian protein disulphide isomerase (PDI) or bacterial DsbC can also be beneficial (Bessette et al., 1999). Although Shewry et al. (1995) reported work in progress to co-express a wheat g-gliadin with plant protein disulphide isomerase (PDI) and binding protein (BiP), this work gave inconsistent results (unpublished results of Sayanova, Napier and Shewry) and was not developed into a routine system. In this respect, it is also worth noting that C hordein of barley forms no disulphide bonds and appears to be readily expressed and purified as a natively folded protein (Tama´s et al., 1994). Eukaryotic yeast and insect cell systems are considered to be more likely to result in correctly folded recombinant proteins with disulphide bond formation occurring if the protein is directed into the cell secretory system. However, this was certainly not the case when a baculovirus system was used to express an LMW subunit, with only about 10% of the expressed protein being recovered as correctly folded (Thompson et al., 1994). There is also no evidence from the literature that wheat gluten proteins expressed in yeast are more authentically folded than those expressed in E. coli, with Scheets and Hedgcoth (1989) and Pratt et al. (1991) both showing that g-gliadin expressed in Saccharomyces comprised a mixture of apparently correctly folded monomers and misfolded oligomers and polymers (see below). However, both studies were performed some time ago and the recent use only of E. coli-based systems means that there are no comparable studies using Pichia or more modern yeast systems. Many plant proteins are processed post-translationally in the cell secretory system (ER, Golgi, vacuole), the most common modifications being glycosylation and proteolysis. However, other modifications such as proline hydroxylation, amidation and phosphorylation may also occur. These modifications may also be essential for biological activity or protein stability (Jenkins et al., 1996), but this is not always the case (Khandekar et al., 2001). Such processing does not take place in E. coli but may occur in eukaryotic expression hosts (yeasts, insect cells). However, the precise enzymes and patterns of modification may differ significantly from those in plant cells. For example, the glycosylation process in yeast is not equivalent to that in plant cells (Sadhukhan et al., 1996) and most proteins that are secreted by yeast cells are hyperglycosylated: this includes proteins which are not normally glycosylated in plants! Similarly, the precise pattern of glycosylation in yeast may not only differ from that in plants but may also vary between yeast species (Cereghino and Cregg, 2000).

263

There is no evidence that gluten proteins show significant levels of glycosylation (see discussion of Shewry (1996) and references therein). Hence, the inability of E. coli to perform glycosylation may be advantageous for the expression of gluten proteins. Although there is no evidence that any of the gluten proteins expressed so far in yeast have been glycosylated (Blechl et al., 1992; Neill et al., 1987; Pratt et al., 1991; Scheets and Hedgcoth, 1989); these reports did not use high yielding secretory systems in which glycosylation would be more likely to occur. Glycosylation may also take place in the cultured insect cells used in baculovirus systems. However, there was no evidence that the LMW subunit protein expressed by Thompson et al. (1994) was glycosylated. Yeast cells also carry out proteolytic processing and may give authentic cleavage of plant signal peptides. However, alternative processing sites may also be recognised as reported by Blechl et al. (1992) for an a-gliadin expressed in yeast. Further proteolytic processing may also occur in yeast but the endogenous yeast proteinases (e.g. Kex2, yapsins) differ in specificity from the legumain proteinases, which process many plant storage protein in the vacuole by cleaving adjacent to asparagine residues (Muntz and Shutov, 2002). Consequently, proteolytic processing may occur at different sites in yeast to those recognised in plants. Gluten proteins are not proteolytically processed subsequent to signal peptide cleavage and so expression in E. coli may again have an advantage over eukaryotic cells in avoiding inauthentic processing. The only difference between the recombinant proteins expressed in E. coli and authentic proteins will therefore be the presence of an additional methionine at the N-terminus of the former, arising from the replacement of the sequence encoding the signal peptide with an ATG initiation codon. 3.3. Ease of purification Purification of proteins from heterologous systems is affected by a range of factors including yield and solubility. Ideally, a high yield of soluble authentically folded protein is required but this is not always possible. We have discussed above how high yields of recombinant proteins often result in the formation of inclusion bodies in E. coli. These insoluble deposits are relatively easy to isolate so incorporation into inclusion bodies may be an attractive strategy if the protein is readily solubilised and refolded (Middelberg, 2002; Sørensen et al., 2003). Another widely used strategy is to fuse the recombinant protein to a ‘tag’ to improve its solubility and folding and to facilitate purification, often by using an affinity matrix (see, for example, Guana et al., 1988; Smith and Johnson, 1988; Tsunoda et al., 2005). Commercial affinity tag systems are available for E. coli, yeast and baculovirus systems and include enzymes to specifically cleave the tag from the purified protein. The only example of the use of a tag to purify a ‘gluten protein’ is the study of Elmorjani et al. (1997) on perfect repeat peptides based on wheat gliadins. The peptides were expressed

264

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

as fusions with thioredoxin but six histidine residues were added downstream to allow purification on a nickel chelation column. The recombinant peptides were then released by treatment with cyanogen bromide or trypsin, to cleave at methionine or arginine, respectively, depending on the construct. The low uptake of fusion technology to facilitate the purification of recombinant gluten proteins is not surprising as they are readily purified by exploiting their solubility properties. In particular, extraction with aqueous alcohol (at room temperature or above and with a reducing agent for proteins containing cysteine residues) followed by salt precipitation is usually sufficient to give a highly enriched preparation which can be further purified by RP-HPLC or other forms of chromatography if required. 3.4. Conclusions It is easy to understand why E. coli has become the system of choice for gluten protein expression as it is simple to use, gives high yields and does not carry out unrequired modifications such as glycosylation and proteolytic processing. It also gives high yields of repetitive proteins and peptides although care must be taken to avoid recombination within gene constructs. Furthermore, although eukaryotic systems should have advantages in giving protein which is authentically folded and disulphide-bonded this has not been the case in the work reported so far. However, it should be noted that the studies using eukaryotic systems were all made over a decade ago and it would perhaps be timely to perform more studies using more advanced systems, particularly in Pichia. 4. Determining the authenticity of expressed proteins It is important to confirm that a recombinant protein is correctly processed and folded, ideally by comparison with the corresponding authentic protein purified from wheat. However, in many cases the corresponding wheat protein is not available. Furthermore, the absence of a clearly defined biological activity (e.g. as an enzyme or receptor) makes it difficult to rule out minor differences in the local conformations of the two proteins. Bearing these limitations in mind, the methods outlined below should be sufficient to confirm that the recombinant and authentic proteins are closely similar if not identical. 4.1. SDS-PAGE The usual first step is SDS-PAGE to confirm that the expressed protein has comparable mobility to the authentic protein. If antibodies are available, identity can be confirmed by western blotting and this will also show whether partially degraded or modified forms are also present. However, SDSPAGE may not be sufficiently sensitive to detect small differences in mass. It should be noted that wheat gluten proteins migrate more slowly on SDS-PAGE than would be predicted based on their

molecular masses (Bunce et al., 1985). The reason for this is not known but it could result from incomplete denaturation of the b-turn rich structures formed by the repetitive sequences present in gluten proteins, resulting in the presence of residual secondary structures. It is therefore difficult to predict the precise migration of gluten proteins in SDS-PAGE based on their coding sequences. 4.2. Mass spectrometry (MS) Matrix assisted laser desorption ionisation-time of flight (MALDI-tof) mass spectrometry (MS) or electrospray (ES)-MS can be used to determine the mass of the expressed protein and compare it with that predicted from the gene sequence. ES-MS is particularly accurate for small proteins and peptides and masses may be identical to those calculated. MALDI-tof MS is less accurate with errors of about 0.1% up to mass 25,000 and higher errors above this (Foti et al., 2000). Nevertheless, MS will certainly demonstrate whether substantial post-translational modification, such as proteolysis or glycosylation, has occurred. 4.3. Protein sequence analysis Identity can be readily confirmed by determining peptide sequences by quadruple-time of flight (Q-TOF) MS. However, these sequences may not include the protein N-terminus and Edman degradation (Koehler, 2003) is usually required to confirm that the N-terminus is correct. Although a range of chemical and immunochemical methods are available to detect glycosylation of proteins separated by electrophoresis, these may give misleading results when applied to gluten proteins (as discussed by Shewry, 1996). If glycosylation is suspected (e.g. based on MS analysis) this should be determined by detailed chemical analysis as described for the HMW subunits by Roels and Delcour (1996). 4.4. Folding and disulphide structure The secondary structure contents of recombinant and authentic proteins can be compared using circular dichroism (CD) or Fourier-transform infra-red (FT-IR) spectroscopy (Tatham and Wellner, 2003). Although neither of these methods will detect subtle differences in structure, a more informative study can be made by performing the analyses at increasing temperatures or in the presence of increasing concentrations of chaotropic agents such as urea or guanidinium chloride. Incorrectly folded proteins would be expected to be less stable and to unfold more rapidly under these conditions. The high contents of intra-chain disulphide bonds in monomeric gliadins, two of which involve adjacent cysteines, means that recombinant proteins may have non-authentic disulphide structures. The disulphide structures of the recombinant and authentic proteins can be compared by comparison of the patterns and sequences of peptides generated by digestion with thermolysin (Mu¨ller and Wieser, 1995, 1997;

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

Mu¨ller et al., 1998) or by analysis of data from mass spectrometry (Caporale et al., 1996, 2005). 5. Application to gluten proteins The expression of gluten proteins in heterologous systems was initiated in 1987 and has now become part of the standard ‘tool kit’ of cereal scientists. Although some of this work has been alluded to in the previous sections on expression systems, the discussion below is focused on the proteins themselves and how the use of recombinant proteins has increased our understanding of gluten protein structure and functionality. Table 2 provides an overview of these studies. 5.1. Gliadins The monomeric a- and g-type gliadins pose a challenge for expression in heterologous hosts as their structures are stabilised by precise patterns of intra-chain disulphide bonds (three in the a-type and four in g-type gliadins). Hence, the protein must either be correctly folded in the host cells or be extracted in a form, which can be readily re-folded in vitro. Because of this requirement, early studies used bakers’ yeast (S. cerevisiae) in the hope that the eukaryotic protein assembly machinery would lead to correct folding and disulphide bond formation. Neill et al. (1987) expressed an a-gliadin in yeast under control of the CYC1 regulatory region. The protein accumulated within the cells at a low level (estimated as 0.1% of the total cellular protein) and when separated on SDS-PAGE gave a single band with similar mobility to authentic a-gliadin from wheat, indicating that signal peptide cleavage occurred. However, the samples were dissolved in Laemmli (1970) sample buffer containing reducing agent and the folding of the protein was not determined. Subsequently, Blechl et al. (1992) developed an improved yeast system for a-gliadin by modifying the ARS1 region of replication with that of the 2m plasmid and developing a new medium to allow selection for the plasmid and growth to high cell density. Sonication with 80% ethanol extracted an enriched fraction, which could be used to purify several hundred micrograms of gliadins per litre culture. The recombinant protein eluted at the same position in RP-HPLC as authentic a-gliadin, indicating that the folding of the two preparations was similar, if not identical. N-terminal amino acid sequencing showed that about two thirds of the protein had the correct N-terminus while a third was processed at a site within the signal peptide three residues from the usual processing site. This emphasises the point made above that processing in yeast cells may not accurately reflect that which occurs in plant cells. Scheets and Hedgcoth (1989) also used yeast but to express a g-gliadin. The complete coding sequence was inserted into a vector using the yeast ADH1 promoter and CYC1 transcription terminator. However, protein expression was low, with a small amount of the total (about 0.5–1.5 ng/l) being found in the culture medium. More success was achieved by Pratt et al. (1991) who expressed a g-gliadin cDNA in yeast using the

265

phosphoglycerate kinase promoter under control of the gal 10 upstream activation sequence and two yeast strains, DBY 747 and the protease deficient BJ2168. Extraction of the freezedried cells with 50% (v/v) aqueous propan-1-ol at 60 8C followed by salt precipitation overnight at 4 8C gave a fraction which was essentially pure and corresponded to a yield of 3–5 mg/l culture (h1–2% of total cell protein). N-terminal amino acid sequencing confirmed that signal peptide cleavage occurred at the correct position but the preparation showed poor solubility in 50% (v/v) propan-1-ol or 1–2% (v/v) acetic acid, which are effective solvents for native g-gliadins. SDS-PAGE under non-reducing and reducing conditions showed that the preparation contained a mixture of disulphide-stabilised oligomers and polymers in addition to monomers, indicating that a substantial proportion of the protein was misfolded. Rosenberg et al. (1993) used the g-gliadin expression plasmid prepared by Pratt et al. (1991) to study the trafficking and deposition in yeast. Most of the protein accumulated in the endoplasmic reticulum with only a small amount being transported to the vacuole. In contrast, when the sequence encoding the repetitive N-terminal domain was deleted all the truncated protein was transported to the vacuole. The authors concluded that the N-terminal domain contained information, which facilitated packing of the protein into dense deposits within the ER. However, it is possible that the structure formed by the protein N-terminus facilitates protein aggregation rather than provides a signal which is recognised by the sorting and trafficking machinery within the secretory pathway. These early attempts to express a-gliadins and g-gliadins used bakers’ yeast and far more sophisticated systems are now available in Pichia. The use of these should allow the yields of correctly folded gliadins to be increased to levels which are sufficient to determine the structures and functional properties of the recombinant proteins. However, more recently E. coli systems have been used to express proteins for biomedical studies. Maruyama et al. (1998) used pET vectors to express a-, g- and u-gliadins and LMW and HMW subunits to compare their allergenicity, while several a-gliadin genes have been expressed in E. coli to study the role of the encoded proteins in coeliac disease (Arentz-Hansen et al., 2000; Mazzeo et al., 2003; Senger et al., 2003, 2005). None of these studies have included conformational analyses of the recombinant proteins but it is unlikely that the a-gliadins, g-gliadins or LMW subunits would have been authentically folded. No other expression studies of u-gliadins have been reported but work on LMW and HMW subunits is discussed below. 5.2. Low molecular weight gliadins Salcedo et al. (1979) described a novel group of gliadin-like proteins which were extracted from flour with chloroform: methanol (2:1, v/v). They comprised about 10 components with Mr of 17,000–19,000 and had similar mobilities to gliadins on electrophoresis at low pH. They also resembled gliadins in their amino acid composition except that they

266

Table 2 Summary of expression systems used for wheat gluten proteins and the characterisation of the recombinant proteins Protein

Expression system

Expression vector

Host cell

Yield/level of expression

Further use

Reference

a-Gliadin

Yeast (Saccharomices cerevisiae) Yeast (Saccharomices cerevisiae) Yeast (Saccharomices cerevisiae) Yeast (Saccharomices cerevisiae)

2m-plasmid derived pAY27 2m-plasmid derived pAY33 pYcDE-2

D1113-10B

0.1% total protein

Neill et al. (1987)

D1113-10B

0.5–4.0 mg/l

XP660-19D(MATd, trpl, ade1, bis2) DBY745 and protease deficient BJ2168

40–80 mg/l

SDS-PAGE, western blotting SDS-PAGE, N-terminal sequencing SDS-PAGE

a- and g-Gliadins

E. coli

pET17xb

BL21(DE3)pLysS

10–30 mg/l

Western blotting, N-terminal sequencing T cell recognition assay

LMW gliadins LMW subunit

E. coli Baculovirus

pET11d PAcYm1

BL21SI Spodoptera frugiperda

22 mg/l 30–50 mg/l

Mixograph, bread making Polymerisation studies

LMW subunits

E. coli

pET11a

JM109/AD494(DE3)

20–30 mg/l

Mixogroph

LMW subunit

E. coli

pTrc99a

BLR21(DE3)pLysS

Not given

LMW subunits

E. coli

pET3a

BL21star(DE3)pLysS

40–100 mg/l

HMW subunit fragment

E. coli

HMW subunit 1Dx2 HMW subunit 1Dx5 modified in length HMW subunits 1Dx2, 1Dx5, 1Dy10, 1Dy12 Chimaeric HMW subunits Synthetic HMW subunits HMW subunit 1Dy12.4t

E. coli E. coli

pET3a pET3a

BL21(DE3)pLysS BL21(DE3)pLysS

z7% total rotein 20–40 mg/l

SDS-PAGE, western blotting SDS-PAGE, western blotting SDS-PAGE, western blotting SDS-PAGE SDS-PAGE, RP-HPLC,

E. coli

pET3a

BL21(DE3)pLysS

10–20% total protein in fermenter

E. coli E. coli E. coli

pET3a pET3a

Subunit 1Dx5 Mr 58,000 peptide

E. coli

pET17b

BL21(DE3)pLysS BL21(DE3)pLysS, BL21(DE3)RIL BL21(DE3)pLysS

20 mg/l

Gliadins, LMW subunit, HMW subunit C hordeins

E. coli

pET21d

BL21(DE3)

5–15% total protein

E. coli

pET3d

30 mg/l

Gliadin- related peptides

E. coli

pEQ2 pET21d2 pET32b

HMW subunit-related peptides

E. coli

pET3d

BL21(DE3)2 JM109 (DE3) M152 HMS174 (DE3)2 BL21(DE3)2 BLR(DE3)pLysS BLR(DE3)pLysS

a-Gliadin g-Gliadin g-Gliadin

pKV49

3–5 mg/l

5% total protein

z5 mg/l 15–20% total protein 200–500 mg/l in fermenter 6–10 mg/l

Mixograph Mixograph SDS-PAGE Mixograph

Scheets and Hedgcoth (1989) Pratt et al. (1991)

Arentz-Hansen et al. (2000), Mazzeo et al. (2003) and Senger et al. (2003, 2005) Clarke et al. (2003) Thompson et al. (1993, 1994) Lee et al. (1999a,b) and Ciaffi et al. (1999) Cloutier et al. (2001) Patacchini et al. (2003) Bartels et al. (1985) Galili (1989) D’Ovidio et al. (1997) Be´ke´s et al. (1994) and Dowd and Be´ke´s (2002) Anderson et al. (1996a) Anderson et al. (1996b) Hassani et al. (2004)

Mixograph, SDS-PAGE, surface properties, MALDI-tof MS Study of allergenicity

Buonocore et al. (1998) and Gilbert et al. (2000)

SDS-PAGE, CD, SAXS, Mixograph SDS-PAGE, UV and FTIR spectroscopy

Tama´s et al. (1994, 1998, 2002) Elmorjani et al. (1997) and Sourice et al. (2003)

SDS-PAGE, western blotting, CD and FT-IR spectroscopy

Feeney et al. (2001) and Wellner et al. (2006)

Maruyama et al. (1998)

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

1.0–1.5% total protein

Blechl et al. (1992)

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

contained lower proportions of glutamateCglutamine (23– 27 mol%) and proline (z9–11 mol%) (Prada et al., 1982). Clarke et al. (2003) isolated a family of four cDNAs and genes encoding these LMW gliadins and expressed one component (LMWGli1111) in E. coli using the pET11d vector. Largescale expression was performed using the salt-inducible BL21SI host strain and over 20 mg of protein was purified from four 1 l cultures. Mixing of 5 mg protein into 2 g base flour showed a reduction in dough strength which was similar to that observed when gliadins were added. In contrast, incorporation of the protein using a reduction and reoxidation procedure resulted in increased mixing time, indicating increased dough strength, but this was accompanied by increased resistance breakdown, indicating reduced dough stability. 5.3. LMW subunits of glutenin Thompson et al. (1994) developed a baculovirus expression system for a LMW subunit gene encoded by the D genome of bread wheat (LMWG-1DI), using the transfer vector pAcYM1. They reasoned that the insoluble LMW subunit protein would substitute effectively for the polyhedrin protein, which is also insoluble and can account for up to 50% of the total protein in the insect cells when the wild type vectors are used (Emery and Bishop, 1987). A shuttle vector containing the LMWG-1DI gene was used to co-infect cultured insect cells with the wild type baculovirus, and recombinant virus expressing the glutenin gene in place of the polyhedrin gene was isolated. This was in turn used to re-infect a separate batch of cultured insect cells, which were harvested after 4 days. The use of the same system with a HMW subunit gene resulted in instability during preparation of the recombinant virus (see below and Madgwick et al., 1992) but this was not observed with the LMW subunit gene. The expressed protein accounted for 25–30% of the extracted protein with 30–50 mg being prepared from 1 l of culture. Furthermore, the plant signal sequence was cleaved correctly and the protein accumulated as dense deposits within the lumen of the endoplasmic reticulum of the insect cells. Hence, the work confirmed the anticipated advantages of the baculovirus system in terms of yield and eukaryotic processing. However, the protein was present in disulphide-bonded polymers and proved difficult to solubilise and purify, even after reduction, and over 90% was irreversibly bound to an ion exchange column (CM cellulose) during purification. The protein eluted from the column showed more typical solubility properties and was refolded in vitro to give a mixture of monomers and polymers with secondary structure contents similar to those of LMW subunits prepared from grain. Despite the low yield, Thompson et al. (1993) used the baculovirus system to study the cysteine residues involved in the formation of inter-chain disulphide bonds. Comparison of the amino acid sequences of a-gliadins, g-gliadins and LMW subunits identified two non-conserved cysteines in the latter, which were proposed to form inter-chain bonds. These cysteines were converted to serines as single and double

267

mutants and the four proteins (three mutants and wild type) were expressed, purified and refolded in vitro. In all cases, the reduced and denatured proteins showed single bands of the expected Mr when separated by SDS-PAGE. When refolded the wild type protein showed traces of a monomeric band of slightly faster mobility than the reduced wild type protein, indicating the presence of intra-chain disulphide bonds. However, most of the protein failed to enter the separating gel and was presumably present as polymers. The two single mutants clearly formed an additional band which was of the mobility expected for a dimer, and this was particularly prominent with mutant 2 in which the remaining ‘unpaired’ cysteine was close to the protein N-terminus. Finally, the double mutant gave mainly monomeric bands, which presumably differed in their patterns of intra-chain disulphide bonds. The results of this study were essentially consistent with those from the direct mapping of disulphide bonds in glutenin polymers (reviewed by Shewry and Tatham, 1997) and from expressing mutant and wild type LMW subunits in an in vitro system composed of wheat germ extract and bean microsomes (Orsi et al., 2001, 2004). Most of the recent studies of LMW subunits have used the pET vectors, which have proved to be so effective for other gluten proteins. In particular, the group of Appels and Morell at CSIRO Plant Industry (Canberra, Australia) have expressed two genes derived from the A genome of the diploid Triticum boeoticum (LMWG-E2 and -E4) and one gene derived from the D genome of T. tauschii (LMWG-16/10) using the pET11a vector (Ciaffi et al., 1999; Lee et al., 1999b). The JM109(DE) host strain was initially used and the LMWG-E2 and E4 proteins were readily purified from extracts made with 70% (v/v) aqueous ethanol. In contrast, the LMWG-16/10 protein formed insoluble inclusion bodies in this host strain but was more readily purified when the trxB mutant AD494(DE3) host strain was used. It is possible that the differences in folding and solubility of the proteins result from subtle differences in the sequences around the seventh and eighth cysteine residues and their effect on disulphide bond formation. Lee et al. (1999a) compared the functional properties of the three recombinant LMW subunits using small-scale Mixograph and extensibility tests. The LMWG-E2 and E4 proteins gave significant increases in dough mixing time and peak resistance with the LMWG-16/10 protein being less effective. Similarly, whereas the two A genome-encoded proteins gave significant increases in extensibility and decreases in maximum resistance, the LMWG-16/10 protein gave a decrease in maximum resistance but no significant effect on extensibility. More recently, Cloutier et al. (2001) have used the pTrc99a vector and BLR21(DE3) pLysS host cells to express an i-type LMW subunit from the overstrong variety Glenlea to confirm the identity of the corresponding protein by western blotting and SDS-PAGE, providing the first evidence for the presence of LMWi subunits in mature wheat grain. Finally, Patacchini et al. (2003) have expressed three LMW subunit genes from durum wheat in E. coli using pET3a and the BL21Star(DE3) pLysS host strain. Two of these genes differed

268

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

in having long (LMW1B) and short (LMW21M18) repetitive domains while the third was a mutant form of LMW1B (called LMW1BK) in which the cysteine residue closest to the protein N-terminus had been substituted by an arginine residue. The proteins were deposited in inclusion bodies, which were prepared by centrifugation and then solubilised in 50% acetonitrile and 0.1% trifluoroacetic acid. Yields of 40–100 mg/l of culture were obtained. 5.4. HMW subunits of glutenin The earliest attempt to express a wheat gluten protein in a heterologous system was reported by Bartels et al. (1985). They used an E. coli system to express a partial cDNA encoding about 500 amino acids of an HMW subunit, as a fusion with b-galactosidase. The expressed proteins were labelled with 35S and detected by SDS-PAGE and immunoprecipitation. Although the expression level was low this work did demonstrate the feasibility of expressing highly repetitive gluten proteins in microbes. Subsequently, Galili (1989) reported the first high yield expression of a wheat gluten protein, which was subunit 1Dx2 in E. coli. The sequence encoding the hydrophobic signal peptide was removed and replaced with an ATG-initiation codon and the sequence encoding the mature protein was inserted into an inducible pET expression vector. The amount of the expressed protein was estimated by densitometric scanning of SDS-PAGE gels to account for 7% of the total E. coli proteins at 18 h after induction of expression. Extraction in the absence of a reducing agent showed the presence of a range of oligomers stabilised by inter-chain disulphide bonds. Subsequent studies have shown that pET-based vectors can be used to express HMW subunits on a routine basis. For example, heterologous expression has been used to confirm the identities of a number of novel HMW subunit genes from bread wheat, from T. timopheevi (a related tetraploid with the genome constitution AAGG) and from a related wild species of Aegilops (Li et al., 2004; Wan et al., 2002, 2005). Because of this ease of use there has been little work on other expression systems for HMW subunits, although Madgwick et al. (1992) did report preliminary studies using a baculovirus system. Satisfactory yields were obtained but recombination within the repetitive sequences occurred during the co-infection to prepare the recombinant virus resulting in the expression of truncated proteins. Subsequent work has used E. coli expression for two purposes: to determine the structures of individual subunits and subunit domains and to determine their functional properties in small-scale test systems.

cells and either shaking flasks or a Biostat-B fermenter. The yields in these systems varied considerably. In small-scale cultures (50 and 200 ml), the expressed proteins accounted for about 10–15% of the total cell proteins and in large-scale cultures (2000 ml) and fermenter cultures, about 1–7%. Furthermore, the x-type subunits (1Dx2, 1Dx5) were expressed at higher levels than y-type (1Dy10, 1Dy12) with the expression of the latter being particularly affected by the medium used (4–5% in ZY, 1–2% in fLB). Addition of glucose to fermenter cultures in ZY medium also resulted in increased expression, up to 10–20% of total bacterial protein. The subunits were almost exclusively located in inclusion bodies which could be prepared and then solubilised in 50% (v/v) aqueous propan-2-ol C10 mM dithiothreitol. Comparison of the effects of recombinant and authentic (i.e. from wheat) subunits on the mixing properties of a base dough using reduction/reoxidation in a 2g Mixograph showed that they were essentially similar, indicating that their functional properties were equivalent. Similar studies of recombinant subunits have been reported by Be´ke´s et al. (1994) (subunits 1Dx5, 1Dx2, 1Dy10, 1Dy12), Hassani et al. (2004) (1Dy12.4t from T. tauschii) and Anderson et al. (1996a) (1Dx5/1Dy10 hybrid proteins). The incorporation studies of Be´ke´s et al. (1994) established a relationship between subunit mass and mixing time (Fig. 1). Furthermore, there was excellent agreement between the results obtained by incorporation of recombinant subunit 1Ax1 into the Mixograph and by expression of the 1Ax1 gene in transgenic wheat (Barro et al., 1997; Popineau et al., 2001). In contrast, expression of the 1Dx5 subunit in transgenic wheat resulted in ‘overstrong’ characteristics which were not revealed by incorporation experiments. Hence incorporation may not always provide a valid test offunction as a prelude to genetic engineering. Anderson et al. (1996a) also produced chimeric proteins in which the N-terminus of subunit 1Dx5 was fused to the repetitive and C-terminal domains of subunit 1Dy10 and vice versa. Dough mixing experiments showed that incorporation of either protein resulted in increased mixing time but that the simultaneous incorporation of both subunits gave a greater increase than either subunit on its own.

5.5. Determination of the functional properties of HMW subunits A number of HMW subunits have been expressed in E. coli to determine their functional properties using the 2g Mixograph and Dowd and Be´ke´s (2002) have reported a protocol for their largescale preparation using the pET3a vector, BL21(DE3) pLysS host

Fig. 1. The relationship between the molecular weights of incorporated HMW glutenin subunits and their effect on mixing time (in seconds), including results for subunit 1By20 purified from wheat. Redrawn from Be´ke´s et al. (1994) with permission.

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

5.6. Exploring HMW subunit structure Although HMW subunits are readily purified from flour, for structural studies these are limited by the complex multidomain structure, which makes it difficult to interpret spectroscopic results. For example, the high a-helical content of the N-terminal domain may dominate the spectra determined by CD spectroscopy in solvents such as trifluoroethanol masking weaker signals from other types of secondary structure (Field et al., 1987). van Dijk et al. (1996) and Gilbert et al. (2000) avoided this problem by expressing recombinant peptides derived from the repetitive domain of subunit 1Dx5. These peptides had masses of 16,802 and about 58,000, respectively, and were used for detailed spectroscopic analyses. The Mr 58,000 peptide has subsequently been used for a range of studies including comparison of its behaviour at an air–water ˝ rnebro et al., 2003), interface with that of the whole protein (O verification of its sequence by direct MALDI-tof MS (Foti et al., 2000) and incorporation into dough, using a 2g Mixograph (Buonocore et al., 1998). For the latter study, a range of mutants with up to four cysteine residues (two each close to the N- and C-termini) were constructed and compared for their ability to become incorporated into dough and modify its properties. Only the peptide containing four cysteines resulted in an increase in dough strength suggesting that steric hindrance may have prevented some of the other cysteines from forming inter-chain disulphide bonds.

269

extraction with 70% (v/v) aqueous ethanol followed by salt precipitation (Tama´s et al., 1994). Conformational studies by circular dichroism spectroscopy and small angle X-ray scattering demonstrated that the recombinant protein was essentially identical in conformation to C hordein purified from barley grain. Tama´s et al. (1994) also expressed a mutant form containing cysteine residues at positions 7 and 236 of the mature protein. This form was present as oligomers and polymers (Fig. 2), which were reduced to monomers by dithiothreitol (Greenfield et al., 1998). Dynamic shear rheology of hydrated solid samples showed that the monomeric wild type protein showed significant elasticity but that this was increased on cross-linking (Greenfield et al., 1998). Hence, even the monomeric protein could potentially affect elasticity when incorporated into dough.

5.7. C hordein as an analogue of glutenin C hordein is a barley seed protein which is related in sequence to the sulphur-poor u-gliadins encoded by the A and D genomes of bread wheat (Hsia and Anderson, 2001; Masoudi-Nejad et al., 2002; Tatham and Shewry, 1995). C hordein consists essentially only of repeated sequences based mainly on an octapeptide motif (Pro-Gln-Gln-Pro-Phe-ProGln-Gln) and is a monomeric protein with no cysteine residues. Cloned genes encoding C hordeins have been available since the late 1980s (Entwistle, 1988; Entwistle et al., 1991; Sayanova et al., 1993), while u-gliadin genes were not cloned until much more recently (Hsia and Anderson, 2001; MasoudiNejad et al., 2002). Hence, C hordein was initially selected as a model for sulphur-poor prolamins and for highly repetitive sequences more generally (Tama´s et al., 1994). Furthermore, the demonstration that the protein was intrinsically elastic and that this property increased when cross-links were introduced (as discussed below) led to its exploitation as a model for glutenin polymer structure. The barley genomic clone l hor 1–17 encodes a C hordein protein of 216 residues including a 20 residue signal peptide. The mature protein comprises an extensive repetitive domain (223 residues) flanked by short non-repetitive sequences at the N- and C-termini. For expression in E. coli, the sequence encoding the signal peptide was removed and replaced with an ATG (methionine) initiation codon and inserted into the pET3d vector. Expression in the BL21(DE3) host strain of E. coli gave about 30 mg of recombinant protein/litre of culture, after

Fig. 2. Expression of wild type and mutant (Ser 7 Cys, Thr 236 Cys) C hordeins in E. coli. The proteins were purified from the E. coli cells and separated under non-reducing conditions. The mutant protein forms a ladder of disulphidestabilised oligomers and polymers while the wild type protein forms a single band. Authors’ unpublished results from the study reported by Greenfield et al. (1998).

270

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

Tama´s et al. (1998) therefore constructed two further mutants containing only one cysteine (Cys 7 or Cys 236) and compared the effects of the four proteins on dough mixing properties when incorporated using a reduction–reoxidation procedure and the 2g Mixograph. Incorporation of the mutant with two cysteine residues resulted in similar positive effects on dough strength as those observed when HMW subunit 1Bx7 was incorporated as a control. In contrast, the two single mutants resulted in decreased dough strength because they acted as chain terminators. Finally, although rheological studies showed that the monomeric wild type protein was intrinsically elastic (Greenfield et al., 1998), this appeared to have a diluting effect (by increasing gliadin:glutenin ratio) when used in dough reconstitution experiments, similar to that observed when gliadins were used in similar studies (Fido et al., 1997). The incorporation of the proteins into the glutenin polymers was monitored by size exclusion HPLC (Tama´s et al., 1998) and multi-stacking PAGE (unpublished results of Tama´s and co-workers). Tama´s and co-workers subsequently used the same approach to express and characterise a wider range of proteins based on C hordein, in which the numbers of N- and C-terminal cysteine residues and the length of the repetitive domain were varied. Their studies have demonstrated the ability of the recombinant proteins to modify dough mixing and baking quality parameters by acting as either chain extenders or chain terminators, depending on the number of cysteine residues, and also demonstrated a relationship between protein monomer size and the magnitude of the effects observed (Howitt et al., 2003; Tama´s et al., 2002). The authors have called the mutant forms of C hordein ‘analogue glutenin’ proteins (ANG) to stress their value as tools to study fundamental aspects of structure–function relationships in glutenin. 5.8. Expression of perfect repeat peptides The effectiveness of E. coli systems for expressing repetitive gluten proteins has allowed the development of model systems based on the expression and analysis of perfect repeat peptides. These allow fundamental aspects of gluten protein structure and properties to be explored without the ‘background noise’ resulting from the mutations which frequently occur in the native proteins. Anderson et al. (1996b) constructed and expressed a series of HMW subunit genes. These encoded proteins comprising 16, 32 and 48 identical copies of the HMW subunit repeat sequence PGQGQQGYYPTSPQQ, flanked by the N- and C-terminal domains of subunit 1Dx5. The protein was expressed in E. coli at 10–20% of the total bacterial protein (15–30 mg/l culture) and characterised by SDS-PAGE. Anderson et al. (1996a) also showed that the synthetic subunits resulted in increased mixing time when incorporated into dough, with the magnitude of the effect being directly related to the number of motifs (Fig. 3). Whereas Anderson et al. (1996b) reconstructed a whole subunit comprising N-terminal, C-terminal and repetitive

Fig. 3. The effects of incorporation of synthetic HMW subunit genes comprising 16, 32 and 48 repeat motifs on the mixing time of dough. Redrawn from Anderson et al. (1996a) with permission.

domains, more recent studies have focused only on the repetitive domains of gliadins and HMW subunits. Elmorjani and co-workers at INRA (Nantes, France) developed a system to express peptides comprising up to 32 copies of the Pro-Gln-Gln (PQQPY) motif present in gliadins, using pET vectors (Elmorjani et al., 1997) and C-terminal translation fusion to E. coli thioredoxin to improve solubility. The fusion protein also contained six histidine residues downstream of the thioredoxin allowing purification by affinity chromatography on a nickel column. They subsequently cleaved the proteins to remove the fusion polypeptides and performed detailed studies on the (PQQPY)8 and (PQQPY)17 peptides using FT-IR spectroscopy. These studies showed that the peptides form a protein network which is rich in b-sheet and stabilised by hydrogen bonds when in the hydrated solid state (Sourice et al., 2003). A similar approach was taken by Feeney et al. (2001) who developed a novel strategy to express peptides based on the hexapeptide and nonapeptide motifs present in the y-type HMW subunits of glutenin. Peptides comprising 113 residues (R3) and 203 residues (R6) were purified (Fig. 4) and shown to have similar CD spectra to subunit 1Dy10 and to the recombinant Mr 58,000 peptide from subunit 1Dx5 discussed above. More recently, Wellner et al. (2006) have used the same strategy to prepare a more extensive series of peptides, all comprising 203 residues (i.e. based on R6) but containing varying numbers of repeat motifs with mutations based mainly on the substitutions which are most frequently observed in the native subunits. Comparison of their conformations during

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

271

Fig. 4. Expression of perfect repeat peptides in E. coli. (A) Schematic structures of the recombinant R3 and R6 peptides. Each has Met Ala at the N-terminus followed by alternating hexapeptides (Pro-Gly-Gln-Gly-Gln-Gln) (light grey) and nonapeptides (Gly-Tyr-Tyr-Pro-Thr-Ser-Leu-Gln-Gln) (dark grey). Each repeat motif is perfect with the exception of a cysteine residue C at position 8 with respect to the N-terminus and -7 with respect to the C-terminus. (B) SDS-PAGE of the purified R3 (track b) and R6 (track c) peptides. Track a shows standard proteins of Mr 12,300, 17,200, 30,000, 42,700, 66,250 and 76,000–78,000. The arrows indicate dimeric forms. Reproduced with permission from Feeney et al. (2001).

hydration (by FT-IR spectroscopy) provided novel information on the relationship between peptide sequence and the ability to form regular structures stabilised by inter-chain hydrogen bonds. 5.9. Conclusions There is no doubt that early attempts to express gluten proteins in heterologous systems were focused more on the development of the technology than on exploiting the opportunities offered to answer questions related to protein structure and functionality. The breakthrough came with the application of pET vectors (Galili, 1989), which allow high-level expression of sequences encoding mature gluten proteins (i.e. after removal of the sequence encoding the signal peptide) in E. coli. This system appears to be particularly suitable for gluten proteins, providing high yields of highly repetitive proteins and peptides. At about the same time the first of a series of small-scale quality testing methods was introduced, the 2g Mixograph (Rath et al., 1990), leading to the development of a range of small-scale tests for mixing, extension testing and baking (reviewed by Be´ke´s et al., 2003). These allow functional tests to be performed using amounts of proteins, which can be produced by small-scale expression in flasks in shaking incubators rather than large-scale fermenters. These smallscale tests have been widely used to evaluate the properties of recombinant proteins.

Finally, several recent studies have focused on producing novel ‘model’ repetitive proteins based on motifs present in gliadins and glutenins. These proteins are too large to be routinely produced by peptide synthesis (although smaller peptides could be synthesised and assembled) and in these cases recombinant expression is the only realistic means of production. 6. Overview It is of interest to look back over the ‘history’ of the heterologous expression of wheat gluten proteins and note how the emphasis has changed. Early studies were focused on using eukaryotic systems (yeast, insect cells) to exploit their machinery of protein folding and processing to achieve the production of authentically folded proteins. The results of these studies were disappointing and the emphasis has now moved to E. coli-based systems and in particular the pET vectors which give high-level expression. These systems have proved to be particularly effective for highly repetitive proteins, including HMW subunits, C hordeins and perfect repeat peptides. This is associated with an increasing interest in using heterologous expression to study fundamental aspects of structure-function relationships in simplified model systems rather than just to produce authentic proteins for structure determination. The E. coli/pET system is clearly ideal for this purpose. However, it should also be noted that eukaryotic systems have been improved immensely since much of the work

272

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

described here was performed. Pichia provides a highly attractive system and has effectively replaced bakers’ yeast while other systems based on filamentous fungi (such as Aspergillus) have also become available. Similarly, several Baculovirus systems are now commercially available which provide more ready access to the non-expert. However, none of these systems show sufficient advantages to be preferred to E. coli for current studies. Acknowledgements Laszlo Tama´s is in receipt of a grant from the Hungarian Scientific Research Fund (OTKA T 046703). Rothamsted Research received grant-aided support from the Biotechnology and Biological Sciences Research Council of the United Kingdom. References Anderson, O.D., Be´ke´s, F., Gras, P.W., Kuhl, J.C., Tam, A., 1996a. Use of a bacterial expression system to study wheat HMW glutenins and the construction of synthetic HMW glutenin genes. In: Wrigley, C.W. (Ed.), Proceedings of the Sixth International Gluten Workshop. Royal Australian Chemical Institute, Cereal Chemistry Division, North Melbourne, Australia, pp. 195–198. Anderson, O.D., Kuhl, J.C., Kuhl, A.T., 1996b. Construction and expression of a synthetic wheat storage protein gene. Gene 174, 51–58. Arentz-Hansen, E.H., McAdam, S.N., Molberg, O., Kristiansen, C., Sollid, L.M., 2000. Production of a panel of recombinant gliadins for the characterisation of T cell reactivity in coeliac disease. Gut 46, 46–51. Aslund, F., Beckwith, J., 1999. The thioredoxin superfamily: redundancy, specificity and gray area genomics. Journal of Bacteriology 181, 1375–1379. Barro, F., Rooke, L., Be´ke´s, F., Gras, P., Tatham, A.S., Fido, R.J., Lazzeri, P., Shewry, P.R., Barcelo, P., 1997. Transformation of wheat with HMW subunit genes results in improved functional properties. Nature Biotechnology 15, 1295–1299. Bartels, D., Thompson, R.D., Rothstein, S., 1985. Synthesis of a wheat storage protein subunit in Escherichia coli using novel expression vectors. Gene 35, 159–167. Be´ke´s, F., Anderson, O.D., Gras, P.W., Gupta, R.B., Tam, A., Wrigley, C.W., Appels, R., 1994. The contribution to mixing properties of 1D glutenin subunits expressed in a bacterial system. In: Henry, R., Ronalds, J. (Eds.), Improvement of Cereal Quality by Genetic Engineering. Plenum Press, New York, USA, pp. 97–104. Be´ke´s, F., Lukow, O.M., Uthayakumaran, S., Mann, G., 2003. Small-scale quality measurements. In: Shewry, P.R., Lookhart, G. (Eds.), Wheat Gluten Protein Analysis. AACC, St Paul, MN, pp. 173–198. Bergkamp, R.J.M., Kool, I.M., Geerse, R.H., Planta, R.J., 1992. Multi-copy integration of the a-galactosidase gene from Cyamopsis tetragonoloba into the ribosomal DNA of Kluyveromyces lactis. Current Genetics 21, 365–370. Bessette, P.H., Aslund, F., Beckwith, J., Georgiou, G., 1999. Efficient folding of proteins with multiple disulfide bonds in the Escherichia coli cytoplasm. Proceedings of the National Academy of Sciences USA 96, 13703–13708. Bhandari, P., Gowrishankar, J., 1997. An Escherichia coli host strain useful for efficient overproduction of cloned gene products with NaCl as the inducer. Journal of Bacteriology 179, 4403–4406. Blechl, A.E., Thrasher, K.S., Vensel, W.H., Greene, F.C., 1992. Purification and characterization of wheat alpha-gliadin synthesized in the yeast, Saccharomyces cerevisiae. Gene 116, 119–127.

Bunce, N.A.C., White, R.P., Shewry, P.R., 1985. Variation in estimates of molecular weights of cereal prolamins by SDS-PAGE. Journal of Cereal Science 3, 131–142. Buonocore, F., Bertini, L., Ronchi, C., Be´ke´s, F., Caporale, C., Lafiandra, D., Gras, P., Tatham, A.S., Greenfield, J., Halford, N.G., Shewry, P.R., 1998. Expression and functional analysis of Mr 58,000 peptides derived from the repetitive domain of HMW subunit 1Dx5. Journal of Cereal Science 27, 209–215. Caporale, C., Sepe, C., Caruso, C., Pucci, P., Buonocore, V., 1996. Assignment of protein disulphides by a computer method using mass spectrometric data. FEBS Letters 393, 241–247. Caporale, C., Bertini, L., Pucci, P., Buonocore, V., Caruso, C., 2005. CysMap and CysJoin: database and tools for protein disulphides localisation. FEBS Letters 579, 3048–3054. Cereghino, C.J.L., Cregg, J.M., 2000. Heterologous protein expression in the methylotrophic yeast Pichia pastoris. FEMS Microbiology Reviews 24, 45–66. Ciaffi, M., Lee, Y.K., Tama´s, L., Gupta, R., Skerritt, H.J., Appels, R., 1999. The low-molecular-weight glutenin subunit proteins of primitive wheats. III. The genes from D-genome species. Theoretical and Applied Genetics 1, 135–148. Clarke, B.C., Phongkham, T., Gianibelli, M.C., Beasley, H., Be´ke´s, F., 2003. The characterisation and mapping of a family of LMW-gliadin genes: effects on dough properties and bread volume. Theoretical and Applied Genetics 106, 629–635. Cloutier, S., Rampitsch, C., Penner, G.A., Lukow, O.M., 2001. Cloning and expression of a LMW-i glutenin gene. Journal of Cereal Science 33, 143–154. Dai, X., Willis, L.G., Palli, S.R., Theilmann, D.A., 2005. Tight transcriptional regulation of foreign genes in insect cells using an ecdysone receptor-based inducible system. Protein Expression and Purification 42, 236–245. Daly, R., Hearn, M.T.W., 2005. Expression of heterologous proteins in Pichia pastoris: a useful experimental tool in protein engineering and production. Journal of Molecular Recognition 18, 119–138. D’Ovidio, R., Anderson, O.D., Masci, S., Skerritt, J., Porceddu, E., 1997. Construction of novel wheat high-Mr glutenin subunit gene variability: modification of the repetitive domain and expression in E. coli. Journal of Cereal Science 25, 1–8. Dowd, C., Be´ke´s, F., 2002. Large-scale expression and purification of HMW glutenin subunits. Protein Expression and Purification 25, 97–104. Elmorjani, K., Thie´vin, M., Michon, T., Popineau, Y., Hallet, J.N., Gueguen, J., 1997. Synthetic genes specifying periodic polymers modelled on the repetitive domain of wheat gliadins: conception and expression. Biochemical and Biophysical Research Communications 239, 240–246. Emery, V.C., Bishop, D.H.L., 1987. The development of multi-expression vectors for high level synthesis of eukaryotic proteins: expression of LCMV-N and AcNPV polyhedrin protein by a recombinant baculovirus. Protein Engineering 1, 359–366. Entwistle, J., 1988. Primary structure of a C-hordein gene from barley. Carlsberg Research Communications 53, 247–258. Entwistle, J., Knudsen, S., Mu¨ller, M., Cameron-Mills, V., 1991. Amber codon suppression: the in vivo and in vitro analysis of two C-hordein genes from barley. Plant Molecular Biology 17, 1217–1231. Faber, K.N., Harder, W., Ab, G., Veenhuis, M., 1995. Methylotrophic yeast as factories for the production of foreign proteins. Yeast 11, 1331–1344. Feeney, K.A., Tatham, A.S., Gilbert, S.M., Fido, R.J., Halford, N.G., Shewry, P.R., 2001. Synthesis, expression and characterisation of peptides comprised of perfect repeat motifs based on a wheat seed storage protein. Biochimica et Biophysica Acta 1546, 346–353. Fido, R.J., Be´ke´s, F., Gras, P.W., Tatham, A.S., 1997. Effects of alpha-, beta-, gamma- and omega-gliadins on the dough mixing properties of wheat flour. Journal of Cereal Science 26, 271–277. Field, J.M., Tatham, A.S., Shewry, P.R., 1987. The structure of a high Mr subunit of durum-wheat (Triticum durum) gluten. Biochemical Journal 247, 215–221. Foti, S., Maccarrone, G., Saletti, R., Roepstorff, P., Gilbert, S., Tatham, A.S., Shewry, P.R., 2000. Verification of the cDNA deduced sequence of

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274 glutenin subunit 1Dx5 and an Mr 58,000 repetitive peptide by matrixassisted laser desorption/ionisation mass spectrometry (MALDI-MS). Journal of Cereal Science 31, 173–183. Galili, G., 1989. Heterologous expression of a wheat high molecular-weight glutenin gene in Escherichia coli. Proceedings of the National Academy of Sciences USA 86, 7756–7760. Gellissen, G., 2000. Heterologous protein production in methylotrophic yeasts. Applied Microbiology and Biotechnology 54, 741–750. Gellissen, G., Melber, K., 1996. Methylotrophic yeast Hansenula potvmorpha as production organism for recombinant pharmaceuticals. Drug Research 46, 943–948. Gellissen, G., Weydemann, U., Strasser, A.W.M., Piontek, M., Janowicz, Z.A., Hollenberg, C.P., 1992. Progress in developing methylotrophic yeasts as expression systems. Trends in Biotechnology 10, 413–417. Gilbert, S.M., Wellner, N., Belton, P.S., Greenfield, J.A., Siligardi, G., Shewry, P.R., Tatham, A.S., 2000. Expression and characterisation of a highly repetitive peptide derived from a wheat seed storage protein. Biochimica et Biophysica Acta 1479, 135–146. Greenfield, J.J.A., Ross-Murphy, S.B., Tama´s, L., Be´ke´s, F., Halford, N.G., Tatham, A.S., Shewry, P.R., 1998. Rheological properties of monomeric and polymeric forms of C hordein, a sulphur-poor prolamin of barley. Journal of Cereal Science 27, 233–236. Guana, C., Lib, P., Riggsa, P.D., Inouyeb, H., 1988. Vectors that facilitate the expression and purification of foreign peptides in Escherichia coli by fusion to maltose-binding protein. Gene 67, 21–30. Hasan, N., Szybalski, W., 1995. Construction of lacits and lacI(q)TS expression plasmids and evaluation of the thermosensitive lac repressor. Gene 163, 35–40. Hassani, M.E., Gianibelli, M.C., Solomon, R.G., Tama´s, L., Shariflou, M.R., Sharp, P.J., 2004. Bacterial expression, purification and functional studies of a y-type high molecular weight glutenin subunit in T. tauschii. In: Lafiandra, D., Masci, S., D’Ovidio, R. (Eds.), The Gluten Proteins. The Royal Society of Chemistry, pp. 58–61. Hinnen, A., Buxton, F., Chaudhuri, B., Heim, J., Hottiger, T., Meyhack, B., Pohlig, G., 1994. Gene expression in recombinant yeast. In: Smith, A. (Ed.), Gene Expression in Recombinant Microorganisms. Marcel Dekker, New York, NY, USA, pp. 121–193. Howitt, C.A., Tama´s, L., Solomon, R., Gras, P.W., Morell, M.K., Be´ke´s, F., Appels, R., 2003. Modifying flour to improve functionality. In: Cauvain, S.P. (Ed.), Bread Making: Improving Quality. CRC Press, Boston, MA, USA, pp. 220–252. Hsia, C.C., Anderson, O.D., 2001. Isolation and characterization of wheat u-gliadin genes. Theoretical and Applied Genetics 103, 37–44. Ikonomou, L., Schneider, Y.-J., Agathos, S.N., 2003. Insect cell culture for industrial production of recombinant proteins. Applied Microbiology and Biotechnology 62, 1–20. Jenkins, N., Parekh, R.B., James, D.C., 1996. Getting the glycosylation right— implications for the biotechnology industry. Nature Biotechnology 14, 975–981. Jonasson, P., Liljeqvist, S., Nygren, P.A., Stahl, S., 2002. Genetic design for facilitated production and recovery of recombinant proteins in Escherichia coli. Biotechnology and Applied Biochemistry 35, 91–105. Khandekar, S.S., Silverman, C., Wells-Marani, J., Bacon, A.M., Birrell, H., Brigham-Burke, M., DeMarini, D.J., Jonak, Z.L., Camilleri, P., FishmanLobell, J., 2001. Determination of carbohydrate structures N-linked to soluble CD154 and characterization of the interactions of CD40 with CD154 expressed in Pichia pastoris and Chinese hamster ovary cells. Protein Expression and Purification 23, 301–310. Koehler, P., 2003. Amino acid and protein sequence analysis. In: Shewry, P.R., Lookhart, G.L. (Eds.), Wheat Gluten Protein Analysis. American Association of Cereal Chemists, St Paul, MN, USA, pp. 137–158. Kost, T.A., Condreay, J.P., 1999. Recombinant baculoviruses as expression vectors for insect and mammalian cells. Current Opinion in Biotechnology 10, 428–433. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. LaVallie, E.R., DiBlasio, E.A., Kovacic, S., Grant, K.L., Schendel, P.F., McCoy, J.M., 1992. A thioredoxin gene fusion expression system that

273

circumvents inclusion body formation in the E. coli cytoplasm. Biotechnology 11, 187–193. Lee, Y.K., Be´ke´s, F., Gras, P., Ciaffi, M., Morell, M.K., Appels, R., 1999a. The low-molecular-weight glutenin subunit proteins of primitive wheats. IV. Functional properties of products from individual genes. Theoretical and Applied Genetics 98, 149–155. Lee, Y.K., Ciaffi, M., Appels, R., Morell, M.K., 1999b. The low-molecularweight glutenin subunit proteins of primitive wheats. II. The genes from A-genome species. Theoretical and Applied Genetics 98, 126–134. Li, W., Wan, Y., Liu, Z., Liu, K., Liu, X., Liu, B., Li, Z., Zhang, X., Dong, Y., Wang, D., 2004. Molecular characterisation of HMW glutenin subunit allele 1Bx14: further insights into the evolution of Glu-B1-1 alleles in wheat and related species. Theoretical and Applied Genetics 109, 1093–1104. Madgwick, P., Pratt, K., Shewry, P.R., 1992. Expression of wheat gluten proteins in heterologous systems. In: Shewry, P.R., Gutteridge, S. (Eds.), Protein Engineering. Cambridge University Press, Cambridge, UK, pp. 188–200. Maruyama, N., Ichise, K., Katsube, T., Kishimoto, T., Kawase, S., Matsumura, Y., Takeuchi, Y., Sawada, T., Utsumi, S., 1998. Identification of major wheat allergens by means of the Escherichia coli expression system. European Journal of Biochemistry 255, 739–745. Masoudi-Nejad, A., Nasuda, S., Kawabe, A., Endo, T.R., 2002. Molecular cloning, sequencing and chromosome mapping of a 1A-encoded u-type prolamin sequence from wheat. Genome 45, 661–669. Mazzeo, M.F., De Giulio, B., Senger, S., Rossi, M., Malorni, A., Siciliano, R.A., 2003. Identification of transglutaminase-mediated deamidation sites in a recombinant a-gliadin by advanced mass-spectrometric methodologies. Protein Science 12, 2434–2442. Middelberg, A., 2002. Preparative protein refolding. Trends in Biotechnology 20, 437–443. Mieschendahl, M., Petri, T., Ha¨nggi, U., 1986. A novel prophage independent trp regulated lambda PL expression system. Biotechnology 4, 802–808. Moffatt, B.A., Studier, F.W., 1987. T7 lysozyme inhibits transcription by T7 RNA-polymerase. Cell 49, 221–227. Mu¨ller, S., Wieser, H., 1995. The location of disulphide bonds in a-type gliadins. Journal of Cereal Science 22, 21–27. Mu¨ller, S., Wieser, H., 1997. The location of disulphide bonds in monomeric g-type gliadins. Journal of Cereal Science 26, 169–176. Mu¨ller, S., Wieser, H., Popineau, Y., 1998. Disulphide bonds of g46-gliadin. Journal of Cereal Science 27, 23–25. Muntz, K., Shutov, A.D., 2002. Legumains and their functions in plants. Trends in Plant Science 7, 340–344. Neill, J.D., Litts, J.C., Anderson, O.D., Greene, F.C., Stiles, J.J., 1987. Expression of a wheat alpha-gliadin gene in Saccharomyces cerevisiae. Gene 55, 303–317. ˝ rnebro, J., Nylander, T., Eliasson, A.-C., Shewry, P.R., Tatham, A.S., O Gilbert, S.M., 2003. The behaviour of the high molecular-weight glutenin subunit 1Dx5, the 58 kDa central repetitive domain and a-gliadins at the air–aqueous interface. Journal of Cereal Science 38, 147–156. Orsi, A., Sparvoli, F., Ceriotti, A., 2001. Role of individual disulfide bonds in the structural maturation of a low molecular weight glutenin subunit. Journal of Biological Chemistry 276, 32322–32329. Orsi, A., Sparvoli, F., Della Cristina, P.A., Ceriotti, A., 2004. Heterotypic aggregation of an in vitro synthesized low-molecular-weight glutenin subunit. In: Lafiandra, D., Masci, S., D’Ovidio, R. (Eds.), The Gluten Proteins. Royal Chemistry Society, Cambridge, UK, pp. 26–29. Patacchini, C., Masci, S., D’Ovidio, R., Lafiandra, D., 2003. Heterologous expression and purification of native and mutated LMW-GS from durum wheat. Journal of Chromatography B 786, 215–220. Philipps, B., Rotmann, D., Wicki, M., Mayr, L.M., Forstner, M., 2005. Time reduction and process optimization of the baculovirus expression system for more efficient recombinant protein production in insect cells. Protein Expression and Purification 42, 211–218. Popineau, Y., Deshayes, G., Lefebvre, J., Fido, R., Tatham, A.S., Shewry, P.R., 2001. Prolamin aggregation, gluten viscoelasticity and mixing properties of transgenic wheat lines expressing 1Ax and 1Dx high molecular weight glutenin subunit transgenes. Journal of Agricultural and Food Chemistry 49, 395–401.

274

L. Tama´s, P.R. Shewry / Journal of Cereal Science 43 (2006) 259–274

Possee, R.D., 1997. Baculoviruses as expression vectors. Current Opinion in Biotechnology 8, 569–572. Prada, J., Sanchez-Monge, R., Salcedo, G., Aragoncillo, C., 1982. Isolation of the major low molecular weight gliadins from wheat. Plant Science Letters 25, 281–289. Pratt, K.A., Madgwick, P.J., Shewry, P.R., 1991. Expression of a wheat gliadin protein in yeast (Saccharomyces cerevisiae). Journal of Cereal Science 14, 223–229. Rath, C.R., Gras, P.W., Wrigley, C.W., Walker, C.E., 1990. Evaluation of dough properties from two grams of flour using the Mixograph principle. Cereal Foods World 35, 572–574. Rietsch, A., Beckwith, J., 1998. The genetics of disulfide bond metabolism. Annual Reviews of Genetics 32, 163–184. Roels, S.P., Delcour, J.A., 1996. Evidence for the non-glycoprotein nature of high molecular weight glutenin subunits of wheat. Journal of Cereal Science 24, 227–239. Rosenberg, N., Shimoni, Y., Altschuler, Y., Levanony, H., Galili, G., Volokita, M., 1993. Wheat (Triticum aestivum L.) g-gliadin accumulates in dense protein bodies within the endoplasmic reticulum of yeast. Plant Physiology 102, 61–69. Sadhukhan, R., Sen, G.C., Sen, I., 1996. Synthesis and cleavage-secretion of enzymatically active rabbit angiotensin-converting enzyme in Pichia pastoris. Journal of Biological Chemistry 271, 18310–18313. Salcedo, G., Prada, J., Aragoncillo, C., 1979. Low MW gliadins-like proteins from wheat endosperm. Phytochemistry 18, 725–727. Sayanova, O.V., Mekhedov, S.L., Zhelnin, L., Khokhlova, T.A., Ananiev, E.V., 1993. Nucleotide sequence of barley C-hordein gene. Genetika 29, 1070–1107. Scheets, K., Hedgcoth, C., 1989. Expression of wheat gamma-gliadin in Saccharomyces cerevisiae from a yeast Adh1 promoter. Journal of Agricultural and Food Chemistry 37, 829–833. Senger, S., Luongo, D., Maurano, F., Mazzeo, M.F., Siciliano, R.A., Gianfrani, C., David, C., Troncone, R., Auricchio, S., Rossi, M., 2003. Intranasal administration of a recombinant a-gliadin down-regulates the immune response to wheat gliadin in DQ8 transgenic mice. Immunology Letters 88, 127–134. Senger, S., Maurano, F., Mazzeo, M.F., Gaita, M., Fierro, O., David, C.S., Troncone, R., Auricchio, S., Siciliano, R.A., Rossi, M., 2005. Identification of immunodominant epitopes of a-gliadin in HLA-DQ8 transgenic mice following oral immunization. Journal of Immunology 175, 8087–8095. Shewry, P.R., 1996. Assembly and post-translation modification of wheat gluten proteins. Gluten’96—Proceedings of the Sixth International Wheat Gluten Workshop. RACI, North Melbourne, Australia (Sydney, Australia, September 1996), pp. 121–128. Shewry, P.R., Tatham, A.S., 1997. Disulphide bonds in wheat gluten proteins. Journal of Cereal Science 25, 207–227. Shewry, P.R., Sayanova, O., Tatham, A.S., Tama´s, L., Turner, M., Richard, G., Hickman, D., Fido, R., Halford, N.G., Greenfield, J., Grimwade, B., Thomson, N., Miles, M., Freedman, R., Napier, J., 1995. Structure, assembly and targeting of wheat storage proteins. Journal of Plant Physiology 145, 620–625. Smith, D.B., Johnson, K.S., 1988. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67, 31–40. Smith, G.E., Summers, M.D., Fraser, M.J., 1983. Production of human beta interferon in insect cells infected with a baculovirus expression vector. Molecular Cell Biology 3, 2156–2165. Sørensen, H.P., Mortensen, K.K., 2005. Advanced genetic strategies for recombinant protein expression in Escherichia coli. Journal of Biotechnology 115, 113–128. Sørensen, H.P., Sperling-Petersen, H.U., Mortensen, K.K., 2003. Dialysis strategies for protein refolding. Preparative streptavidin production. Protein Expression and Purification 32, 252–259.

Sourice, S., Nisole, A., Gue´guen, J., Popineau, Y., Elmorjani, K., 2003. High microbial production and characterization of strictly periodic polymers modelled on the repetitive domain of wheat gliadins. Biochemical and Biophysical Research Communications 312, 989–996. Studier, F.W., Moffatt, B.A., 1986. Use of bacteriophage-T7 RNA-polymerase to direct selective high-level expression of cloned genes. Journal of Molecular Biology 189, 113–130. Studier, F.W., Rosenberg, A.H., Dunn, J.J., Dubendorff, J.W., 1990. Use of T7 RNA polymerase to direct expression of cloned genes. Methods in Enzymology 185, 60–89. Tama´s, L., Greenfield, J., Halford, N.G., Tatham, A.S., Shewry, P.R., 1994. A b-turn rich barley seed protein is correctly folded in Escherichia coli. Protein Expression and Purification 5, 357–363. Tama´s, L., Be´ke´s, F., Greenfield, J., Tatham, A.S., Gras, P.W., Shewry, P.R., Appels, R., 1998. Heterologous expression and dough mixing studies of wild-type and mutant C hordeins. Journal of Cereal Science 27, 15–22. Tama´s, L., Gras, P.W., Solomon, R.G., Morell, M.K., Appels, R., Be´ke´s, F., 2002. Chain extension and termination as a function of cysteine content and the length of the central repetitive domain in storage proteins. Journal of Cereal Science 36, 313–325. Tatham, A.S., Shewry, P.R., 1995. The S-poor prolamins of wheat, barley and rye. Journal of Cereal Science 22, 1–16. Tatham, A.S., Wellner, N., 2003. Spectroscopic analysis of gluten protein structure: circular dichroism and infra-red. In: Shewry, P.R., Lookhart, G.L. (Eds.), Wheat Gluten Protein Analysis. American Association of Cereal Chemists, St Paul, MN, USA, pp. 159–172. Thompson, S., Tatham, A.S., Shewry, P.R., 1993. Exploring disulphide bond formation in a LMW subunit of glutenin using a baculovirus expression system. Gluten Proteins. Association of Cereal Research, Detmold, Germany, pp. 345–355. Thompson, S., Bishop, D.H.L., Madgwick, P., Tatham, A.S., Shewry, P.R., 1994. High-level expression of a wheat LMW glutenin subunit using a baculovirus system. Journal of Agricultural and Food Chemistry 42, 426–431. Tsunoda, Y., Sakai, N., Kikuchi, K., Katoh, S., Akagi, K., Miura-Ohnuma, J., Tashiro, Y., Murata, K., Shibuya, N., Katoh, E., 2005. Improving expression and solubility of rice proteins produced as fusion proteins in Escherichia coli. Protein Expression and Purification 42, 268–277. van den Berg, B., Ellis, R.J., Dobson, C.M., 1999. Effects of macromolecular crowding on protein folding and aggregation. EMBO Journal 18, 6927–6933. van Dijk, A.A., DeBoef, E., Bekkers, A., van Wijk, L.L., van Swieten, E., Hamer, R.J., Robillard, G.T., 1996. Structural characterisation of the central repetitive domain of high molecular weight gluten proteins. 2. Characterisation in solution and the dry state. Protein Science 6, 649–656. Vasina, J.A., Baneyx, F., 1997. Expression of aggregation-prone recombinant proteins at low temperatures: a comparative study of the Escherichia coli cspA and tac promoter systems. Protein Expression and Purification 9, 211–218. Wan, Y., Wang, D., Shewry, P.R., Halford, N.G., 2002. Isolation and characterisation of five novel high molecular weight subunit genes from Triticum timopheevi and Aegilops cylindrial. Theoretical and Applied Genetics 104, 828–839. Wan, Y., Yan, Z., Liu, K., Zheng, Y., D’Ovidio, R., Shewry, P.R., Halford, N.G., Wang, D., 2005. Comparative analysis of the D genomeencoded high molecular weight subunits of glutenin. Theoretical and Applied Genetics 111, 1183–1190. Wellner, N., Marsh, J., Savage, A., Halford, N.G., Shewry, P.R., Mills, E.N.C., Belton, P.S., in press. Structure–function relationships in an elastomeric protein of wheat: synthesis, expression and characterisation of repetitive peptides based on sequence variants present in the HMW subunits. Biomacromolecules. Zapun, A., Missiakas, D., Raina, S., Creighton, T.E., 1995. Structural and functional-characterization of DsbC, a protein involved in disulfide bond formation in Escherichia coli. Biochemistry 34, 5075–5089.