VOLUME IV

INDUSTRIAL PRODUCTION WITH IMMOBILIZED ENZYMES: HIGH-FRUCTOSE CORN SYRUPS AND AMINO ACIDS FOR FOOD AND FEED

(Flowsheets of the processes and other drawings are not shown on this site. Please contact the author, if needed)

INDUSTRIAL ENZYME ENGINEERING:
PRESENT STATUS AND FUTURE PROSPECTS

400. The resultant HFCS is the functional equivalent of sucrose for many applications. It is widely used as a nutritional sweetener in the food industry, particularly in soft drinks and other beverages, as well as in desserts, baked goods, canning, and many other packaged products. For certain uses (e.g., cola beverages), HFCS with 55% or 90% fructose is preferred, made from the 42%-fructose HFCS by a process of separation analogous to liquid chromatography (Carasik & Carroll, 1983).

401. The rapid growth in demand for fructose in the last decade may be attributed to the following:
       - it can be used as an alternative to sugar since both are equally sweet,
       - it can be produced at a cost lower than sugar,
       - it is produced from starch, which is a common and widely available food material,
       - its taste is more refreshing than that of sugar,
       - consumption of fructose is said to be of lower risk for diabetics and some other metabolic
         disorders.

402. The amount of fructose produced by the isomerase reaction is determined by the equilibrium value of the glucose to fructose isomerization which results in roughly equal amounts of the two sugars. In order to obtain higher fructose content of the syrups, large-scale columns of strong acid-ion exchange in the form of calcium salts are often used as the stationary phase. Oligosaccharides come off the columns first, owing to a molecular exclusion effect. These are then followed by glucose and finally fructose.

403. Another alternative is to concentrate the syrup to a high enough content of solids to allow crystallization of glucose from the concentrate. The remaining liquid is then enriched in fructose. In this manner it has been possible to make syrups of up to 70% fructose (on a dry basis). The isolated glucose is recycled in an isomerization reaction. Finally, another approach is to complex fructose with borate compounds. This removes fructose from solutions and shifts the apparent equilibrium of the reaction. However, the cost of removal and recovery of the borate has so far prevented commercial adaptation of this process (Hultin, 1983).

2.     Commercial preparations of immobilized glucose isomerase

404. Clinton Corn Processing Co. developed the first industrial process HFCS production, which used the microorganisms Streptomyces sp. as the source of glucose isomerase. In contrast, Novo Laboratories Inc. selected the thermophilic microorganism Bacillus coagulans, which produces glucose isomerase without requiring the addition of an expensive inducer such as xylose in the mixture (Carasik & Carroll, 1983), and can be grown in continuous culture without contamination and with high enzyme yields. In spite of the large number of organisms that can produce glucose isomerase (Antrim et al., 1979), only a few high-producing strains are actually being utilized on a commercial basis.

405. Glucose isomerase typically is an intracellular enzyme. The enzyme is generally obtained by scaling up a submerged aerated fermentation in several stages. The three stages of development for such a production are described in U.S. Patent 3,666,628: (a) slant development, (b) culture development - two substages, (c) final fermentation stage. Glucose isomerase can be recovered in two ways: (i) recovery of the microorganisms with the enzyme entrapped in the cellular mass, and (ii) recovery of the enzyme in the soluble form after lysing the cells (Antrim et al., 1979). To prevent the loss of the enzyme from cells in the first case (as the result of the autolysis of cells during isomerization) and to allow extended use of cells in fixed-bed reactors, the enzyme is bound to the cellular matrix by heat or chemical fixation (Antrim et al., 1979).

406. Commercial immobilized glucose isomerase preparations are generally produced by various companies in a granular form and a fibrous or amorphous form. Clinton Corn Processing Company, for example, produces both a fibrous form and a granular form of glucose isomerase. Each form is designed for a particular reactor technology. The fibrous form of Clinton's enzyme preparation has a large surface area with very high enzyme potency and is designed for use in shallow-bed reactors. The granular enzyme with lower potency is designed for deep-bed reactors. Novo Enzyme Corporation produces a granular isomerase suitable for deep-bed reactors, batch isomerization, and fluidized-bed operations. The immobilized enzyme preparation from Gist Brocades (Maxazyme^®) utilizes the organism A. missouriensis entrapped in cross-linked gelatin. This produces a softer particulate enzyme which can also be used in deep-bed reactors.

407. ICI produces an immobilized enzyme preparation whereby Arthrobacter cells are flocculated by polyionic reagents. This also produces a rather soft granular particle that can be used in deep-bed reactors. Sanmatsu (Japan) produces a glucose isomerase by adsorption of the enzyme on an anion-exchange resin. This yields a high-potency, particulate enzyme suitable for deep-bed reactors. Denki-Kagaku - Nagase Sangyo (Japan) has produced a polymer isomerase (Sweetase^®) entrapped in hard granules for use in deep-bed reactors. The catalyst particles swell to about double their original sizes (0.4-0.8 mm) during the isomerization reaction (Okada, 1978).

408. Miles and Miles-Kali Chemie have produced a glutaraldehyde cross-linked preparation and a heat-fixed cellular preparation, respectively. Snamprogetti (Italy) produces glucose isomerase, extracted and partially purified from Streptomyces sp., and entrapped in cellulose triacetate fibers. This is carried out according to Snamprogetti's modification of the well-known technology of fiber wet spinning: an organic solution of fiber-forming polymer is emulsified with an aqueous solution of the enzyme, and the resulting emulsion is then extruded into a coagulation bath through the holes of a spinneret (Marconi & Morisi, 1979).

409. The main conclusion from the above table is that there are actually no covalently bonded enzymes among commercial preparations of immobilized glucose isomerase. Glucose isomerase is either adsorbed to ion exchange resins or to porous inorganic carriers (alumina, ceramics), or it is inside whole cells entrapped in a polymer matrix, usually cross-linked with glutaraldehyde or other bifunctional reagents. Even Corning Glass Works, which has developed a well-known procedure for covalent immobilization of enzymes on inorganic carriers, uses immobilization by adsorption in this particular case. The main reason for this is that immobilized glucose isomerases that have been produced by adsorption of soluble enzyme onto a solid carrier can generally be regenerated and reloaded with fresh enzyme after much of the initial activity is depleted. It is usually economically advantageous to do so, especially when the carrier used is expensive.

410. In many cases whole cells (heat-fixed or entrapped into polymer matrix) are used instead of the isolated enzyme. This can be explained both by decreased stability of isolated (extracellular) glucose isomerase (Ishikawa, 1977) and by the higher cost of isolation and immobilization of enzymes in comparison with the use of whole microbial cells.

411. These two points (par. 409 and 410) are illustrated below in the form of a comparison of the general procedures of enzyme immobilization with respect to several important commercial properties (Antrim et al., 1979).

412. The increase in HFCS penetration of the world sweetener market has been made possible in part by the evolution of immobilized glucose isomerase technology and improvements in enzyme characteristics. The dynamics can be traced using the example of Novo Laboratories Inc., which subdivided various types of immobilized glucose isomerase preparations developed by the company into the preparation of three generations of enzymes (Carasik & Carroll, 1983). These do not include the first of a series of commercial glucose isomerases marketed under the brand name Sweetzyme^®, which was a spray-dried soluble product introduced in 1973. It had a high production cost and was used commercially for only a short time. When Novo's first immobilized form of the enzyme was introduced in 1974, its advantages became readily apparent and use of the spray-dried soluble form was subsequently discontinued (Carasik & Carroll, 1983).

413. The steps in the production of immobilized glucose isomerase of "three generations" (and for three types of reactors) are shown in Figure 1 and described below (Carasik & Carroll, 1983).

414. For production of the first generation enzyme (1974), in the form of a powder, the cross-linking agent selected is glutaraldehyde, which immobilized the cell protein. The initial product of cross-linking is a semigelatinous mass that has the consistency of pudding. After drying, grinding and screening, the final product has a particle size range of 100-350 microns. The product is designed for use (and re-use) in a batch reactor (Fig. 2).

415. The first-generation immobilized enzyme has several limitations. The particles gradually decrease in size as a result of the shearing effect of the agitator blades. This leads to an overall decrease in settling rate and thereby to a slowing of the process. According to Novo's data, decrease in the size of the particles of the immobilized enzyme from 0.35 to 0.25 mm and further to 0.10 mm leads to a decrease in the terminal velocities of settling particles from 9.5 to 5.0 m/hr and further to 0.8 m/hr. In addition to the degradation of the particle, there is thermal denaturation of the enzyme, decreasing its activity. Prolongation of the reaction time or addition of new enzyme to the isomerization reaction tank prior to the next cycle can compensate for this. Normal make-up is about 5% per reuse. Finally, the long holding time required for the isomerization with the first generation enzyme in turn requires operation at a relatively low pH to avoid excessive color and formation of by-products. At this pH, the addition of cobalt salts is necessary to activate the immobilized enzyme. These drawbacks of the powder form of the enzyme and the batch reactor led to the development of immobilized enzymes of the second generation, which are suitable for continuous column reactors, where the reaction can be run at a neutral to slightly alkaline pH, since the syrup contact time is reduced drastically, with no cobalt addition required.

416. The second-generation immobilized enzyme for column use was prepared by the extrusion of the wet cross-linked enzyme mass through a small orifice (< 1 mm diameter) cut into short cylindrical particles, which are then dried and screened, as shown in Figure 1. This product does not exhibit pressure-drop characteristics; upon continued operation the particles deform and the pressure drops dramatically. This is overcome by modifying the immobilized enzyme preparation through the incorporation of powdered alumina to increase the density of the particles and allow higher flow rates in the upflow mode (Carasik & Carroll, 1983). However, industrial performance is entirely satisfactory, and the deviation from laboratory data is ascribed to variations in enzyme particle size and to problems of density and flow-distribution problems.

417. For the preparation of the third-generation immobilized glucose isomerase, which should be able to resist deformation at constant syrup flow for fixed-bed operation, an important change in the immobilization process is made. The cell slurry is homogenized prior to cross-linking as shown in Figure 1. Homogenization disrupts the cell membranes and allows more of the protein surface to react with the cross-linking agent. The result is an enzyme with stronger particles that resist deformation. The improved enzyme preparation also contains additives such as magnesium oxide and dextrose, which are incorporated after crosslinking but prior to extrusion. Magnesium oxide helps to minimize the decrease in pH during isomerization, and the dextrose dissolves during operations, resulting in a more porous matrix and lowering diffusion resistance (Carasik & Carroll, 1983). The enzyme is shipped to the user in a dry granular form which, prior to use, must be hydrated in syrup. During hydration, the size of the enzyme particles expands by a factor of about two. Under operating conditions (pH 7.5-8.0, temperature 55-60°) the half-life of the enzyme is usually greater than 75 days.

3.     Technological characteristics of the processes

418. The literature contains limited information on commercially available glucose isomerase technologies. Listed below are some data on reactor configurations normally employed, and basic performance data where available. Although each system listed employs a different enzyme source (see above), a different level of enzyme purity, and a different immobilization technique, there appears to be relatively little difference in the performance of the systems for which data are available (Sweigart, 1978).

419. Major commercial manufacturers of immobilized glucose isomerase describe the application for deep-bed reactors. A deep-bed reactor is usually simpler in design than a shallow-bed model and may therefore require a lower capital expenditure. Flow may be "up" or "down" through the reactor. Column reactors operating in the "down" mode appear to be more popular. Heights up to five meters appear to be feasible. Upflow in a column is reported to be less productive, probably due to wider distribution of the "syrup residence time." Precautions must be taken with upflow to prevent loss of immobilized enzyme through the top of the reactor, and flow control is considered to be more critical than for downflow (Antrim et al., 1979). The temperature and pH recommended are almost identical (60-65°C and 7.0-8.5 respectively). Further, the initial substrate of 40-50% dry solids (92-94% dextrose, dry basis, 6-8% dry basis polysaccharides) is also much the same.

420. In continuous production of isomerized sugar by means of the immobilized enzyme process, raw material purity influences catalyst activity and therefore productivity. Thus, raw material (dextrose syrups) should be refined as much as possible. Impurities include calcium ions, peptides, oxygen, oxidation products, etc. According to the data of the Chiba Plant, Denki Kagaku Kogyo K.K. (Okada, 1978), when high purity substrates such as crystalline glucose are used as the initial substrate, the productivity of the reactor (i.e., the yield of solid high fructose syrup per unit weight of catalyst until catalyst activity decreases to a quarter of the original level) reaches 4000 kg/kg of immobilized enzyme in 100 days, and the half-life for the catalyst is 50 days. However, when purified dextrose of lower grade is used, the half-life decreases to 20 days, and productivity of the reactor to 1,500 kg fructose/kg of immobilized enzyme. In general, productivity for immobilized glucose isomerases used commercially, described by various enzyme manufacturers and used under the operating conditions recommended by them, varies from 1 to 9 tons of HFCS (dry basis) per kg of the immobilized enzyme.

421. A multiplicity of reactor designs has been described for use with immobilized glucose isomerase (six different design types, according to one particular classification, i.e. batch, packed-bed, continuous-flow stirred-tank, continuous-flow stirred-tank/ultrafiltration membrane, and others, including recycle reactors and tubular reactors with enzymatically active walls) (Antrim et al., 1979). However, most glucose isomerase reactors now in commercial operation are of the packed-bed type. Novo Industri has described batch reuse of Type A, a glutaraldehyde cross-linked homogenate of B. coagulans. Although they subsequently found performance advantages with a continuous system, in 1976 they reported that Type A had been used commercially in large-scale batch reuse since 1974 [cit. in Antrim et al. (1979)]. Gist-Brocades has also described conditions for reuse of their Maxazyme GI-Immob in a batch reactor.

422. Numerous investigators have compared performance and economics of batch versus continuous reactor systems. Usually in a packed-bed reactor the concentration of active glucose isomerase is high compared to a batch reactor and contact time between substrate and enzyme is relatively short, usually 2-4 hr as compared with 20-60 hr for a batch reactor (Antrim et al., 1979). The short contact time helps to minimize formation of colored materials and nonfructose isomerization compounds. Additionally, enzyme usage (comparative amounts of enzyme activity), is found to be considerably higher (1.4-4.0 times) for batch versus continuous packed-bed reactors, primarily due to loss of active enzyme through multiple batch recovery operations (Antrim et al., 1979).

423. U.S. Patents 3,847,740 and 3,847,741 disclose a process for regulating production by means of temperature control as well as for increasing the productivity of the enzyme. One example demonstrates that by increasing the temperature in the reactor by 2°C increments from 60° to 70°C over a period of 14 days, the enzyme productivity increases 42% over a 14-day isothermal run at 60°C.

424. Novo Laboratories Inc. used the 'first-generation' immobilized glucose isomerase (see par. 413) in a batch reactor (Fig. 2). Typically the enzyme is added to a tank containing purified high-dextrose syrup (> 93%, dry basis) adjusted to certain conditions (pH 7.0, temperature 60-65°C, Co⁺² 3.5x10^-4M, MgSO₄x7H₂O 0.1-1.0 g/liter). The typical reaction time is 20-24 hr, after which the enzyme is recovered by allowing the particles to settle and the supernatant liquor drawn off. The entire process is then restarted. The limitations of this mode of process operation are described in paragraph 415.

425. Glucose isomerase of the second generation, i.e. crosslinked with glutaraldehyde and extruded (see par. 416), has led to non-satisfactory pressure-drop characteristics of the column reactor in the downflow operation mode. Ultimately, the entire enzyme bed will compress to a point where flow is reduced drastically and the operation is stopped. Therefore, Novo Labs at that stage of development (1975) attempted to use an expanded- or fluidized-bed reactor, in which the dextrose syrup flows into the bottom of the column through a flow distributor and the immobilized enzyme is suspended in the flowing syrup (Carasik & Carroll, 1983). In the fluidized-bed operation, the substrate solution passes upward through the enzyme bed. In this mode the most critical factor is uniform flow distribution. In commercial reactors this is achieved with a shallow bed of heavy inert material (alumina) on a screen.

426. With third-generation immobilized glucose isomerases, which have stronger particles that resist deformation (par. 417), after the hydration the enzyme slurry is transferred to the column reactor, which runs in an upflow (fluid-bed) mode for about 12-24 hr. This permits the enzyme to come into equilibrium with syrup at column operating conditions without compacting. After this "fluid" period, the bed is allowed to settle briefly, then fluid is run in the downflow mode. This process is currently in wide industrial use in many parts of the world (Carasik & Carroll, 1983). The enzyme bed height at optimum operating conditions is 5 m maximum, diameter 1.5 m maximum (height:diameter ratio is 3:1 minimum), temperature 55-60°C at inlet, pH 7.5-8.0 at 25°C, dextrose concentration at inlet 94% minimum (dry basis), feed solids 45%, Mg⁺² content equals 20x Ca⁺². Half-lives are usually more than 75 days.

427. To achieve high enzyme productivity (in this particular case above 4,000 kg of dry solids/kg of enzyme, at 45% conversion of dextrose to fructose, see also paragraph 420) on a continuing basis, Novo recommends (Carasik & Carroll, 1983) that careful attention be paid to details of plant design and operation. Thus, in reactor design, flow distribution is important for operation over a wide range of flow rates. The enzyme is often used until the flow rate (activity) falls to 10% of the initial flow. Feed purity is of great importance to avoid cumulative enzyme poisoning (cf. paragraph 420). Filtration, decolorization (active carbon treatment), deionization, and evaporation must be designed and operated carefully. It is important to provide a purified feed of constant composition with consistent control of parameters.

428. Data on the isomerization process are also available from Kyowa Hakko Kogyo Co., Ltd (Yokote et al., 1975). Glucose isomerase used in their process was immobilized on a phenol formaldehyde resin Duolite A7. The flow chart of the process is shown in Figure 3. A 40% (w/v) glucose solution at 60°C is fed continuously through the immobilized enzyme columns in series at 60°C, pH 8.2. Before feeding, the substrate solution is passed through pretreatment columns that contain Duolite A7 (buffered) to eliminate some impurities and air bubbles. The effluent solution from the immobilized enzyme column is passed successively through a cation exchange (Diaion SK-1A, H-type) and an anion exchange (Diaion WA20, OH-type) column in order to eliminate salts. Microbial contamination was not observed throughout the whole operation period due most likely to the high glucose concentration (40%) and the high operating temperature (60°C). The half-life of the enzyme catalyst under their operating conditions is 34-40 days, depending on the flow rates. According to the company's data, 1 liter of the immobilized enzyme (an adsorption type) isomerizes 288 kg of glucose within 30 days. On the other hand, 1 liter of a covalently immobilized enzyme (with triazinyl chloride as a coupling agent) isomerizes 576 kg of glucose in the same period.

429. In Clinton Corn Processing's approach (flow sheet shown in Figure 4), the immobilized enzyme constitutes a thin bed of 2.5-12.5 cm depth (Davis, 1974). Several of these beds are staged to form a multibed processing unit resembling a pressure leaf filter (see also paragraph 419). A major advantage of having more than one bed is to minimize the effect of channeling that easily occurs in shallow beds. For bed dimensions, Clinton Corn recommends a depth-to-width ratio of about 0.02 to 0.05 (Antrim et al., 1979; Davis, 1974). Single beds can be removed for regeneration during processing without interruption of the process flow. Carbon treatment and ion exchange are used to purify the 42%-fructose effluent, and it is concentrated to syrup by evaporation. Enzyme half-life, according to Clinton Corn, is "several hundred hours" (Davis, 1974).

430. In the mid 1970s Sanmatsu Kogyo was the world's first manufacturer of sugar obtained by isomerization with soluble glucose isomerase. Mitsubishi Chemical Industries Ltd and Seikagaku Kogyo jointly developed an ion exchange resin that selectively adsorbs glucose isomerase "without affecting its activity" (Ishikawa, 1977). These three companies signed a license contract for this process that led to a commercial test plant that was put into operation in 1975. On the basis of satisfactory test results, Sanmatsu Kogyo decided to switch its whole production process to this system in 1976. Later on that year Sanmatsu factories in Chiba and Fukuoka were remodelled to utilize Mitsubishi's technology. Since then these factories have operated commercially. The production process is summarized as follows (Ishikawa, 1977). The microbial product of glucose isomerase is filtered for culture liquor separation, repulped in water and treated with additives for several hours to extract the enzyme. After separation of the residual cells, the enzyme liquor is allowed to contact the ion exchange resin, which adsorbs almost all the glucose isomerase within several hours. Since the carrier selectively adsorbs glucose isomerase, refining the enzyme liquor is unnecessary. The isomerization itself takes place in a simple column packed with the immobilized enzyme, glucose concentration of 45-50%, reaction temperature 55-70°C, pH 6.5-8.5, space velocity (initial) 1.5-4 hr^-1. The glucose conversion level is 45%. At these operating conditions the half-life of the immobilized enzyme is 50 days if crystalline glucose is used as raw material, but as low as 30 days if the raw material is a saccharified glucose solution subjected to conventional purification. The isomerized glucose syrup is almost colorless and has few byproducts. Simple treatment of the syrup with ion exchange resin completely desalts and decolorizes the product. After being used the immobilized enzyme is washed away from its carrier in the reactor through a regeneration step employing "some chemical solutions". The regenerated carrier is reused for isomerization after adsorbing new, active enzyme.

431. According to Zhang (1982), Chinese biotechnologists used glucose isomerase from Streptomyces roseoruber, adsorbed on a strong basic anion exchange resin 290 (made by Nankai University, China), for making HFCS. Pilot plant scale experiments have been carried out with 1.25-2.2 tons of HFCS produced per kg of dry immobilized enzyme. A general overview of biotechnology programs in China is given in Mang (1991) and Han Ying-Shan (1991).

432. In the Snamprogetti (Italy) process glucose isomerase entrapped in fibers is used for the isomerization of industrial glucose solutions. Under operating conditions (65°C, 50% w/w glucose syrup) the half-life of the immobilized enzyme preparations is about 70 days. The product is colorless, and activated carbon treatment is not required. The best process performances were achieved with tubular and radial reactors (Marconi & Morisi, 1979). In the tubular reactor the packing is obtained with fibers placed in an ordered manner parallel to the long axis of the column, a packing degree in the range of 0.2-0.25 kg of dry fibers per liter of reactor volume results in a stable catalyst bed practically incompressible and with negligible channelling and relatively low pressure drop. On the other hand, the radial reactor fits very well with the filamentous structure of the fibers. It is prepared by rolling the fibers around a perforated pipe in an ordered manner, as on a bobbin; the reactor mixture flows from the holes of the central pipe, passes through the fiber layer where the enzymatic reaction takes place, and exits. Packing degrees up to 0.35 kg of dry fibers per liter of reactor volume are reached, making it possible to increase the efficiency of the fiber-entrapped glucose isomerase and to decrease the residence time. According to Snamprogetti, the radial reactors have the great advantage that they can be easily prepared with standard equipment of the textile industry, such as winding machines. The total productivity during two half-lives of the immobilized enzyme in the reactors is around 5-6 tons of HFCS (dry mass) per kilogram of fibers (Marconi & Morisi, 1979).

433. Figure 5 shows a flow sheet for the Hungarian industrial process with a daily processing capacity of 400 tons of native corn, or 120,000 tons/year (Hollo et al., 1984). The plant based on the wet milling of maize was built in 1983 in Szabadegyhaza, Hungary, and has been operating since then. The first step of the process is the continuous enzymatic liquefaction of the starch slurry, based on the technology of Miles/DDS-Kroer and using thermostable alpha-amylase Optitherm^® from Miles Kali-Chemie. The hydrolysate leaves the dextrinizing tank with a DE value of 15. Liquefaction is followed by the enzymatic saccharification at 60°C and pH 4.5 in batch mixing tanks of 200 m³. For this step, the Aspergillus niger glucoamylase of Miles Kali Chemie (Optidex^®) is used. After a saccharification time of 60 hr, the DE value of the resulting glucose solution is 96-98. The glucose solution is purified by separating the germ oil by rotary drum filtration in the presence of active carbon and by full ionization with ion exchange resins. Following the evaporation step, pH and temperature correction, and the addition of magnesium ions the solution is pumped for continuous isomerization into the columns, packed with immobilized glucose isomerase Takasweet^®, Miles-USA. A recent general overview of biotechnology in Hungary is given in Dibner & Burrill (1988) and Wagner & Groo (1992).

434. In the Hungarian process two columns are fed by down-flow and operated in parallel; new columns are put into operation at the necessary time, when the activity of the immobilized glucose isomerase decreases below the control level. The reaction process proceeds at a temperature of 60-62°C, pH 7.8, and yields 2.5 tons of the syrup (dry mass) per 1 kg of the immobilized enzyme; the enzyme half-life is 40 days. The bed volume of enzyme catalyst in the column is 20 m³, the overall consumption of the immobilized enzyme equals 35 tons/year (Hollo et al., 1983). The isomerization is followed by ion exchange, filtration with active carbon and finally concentration in a 4-stage falling film evaporator to 71% DS with a standard composition of 42% fructose, 52% glucose and 6% malto-oligosaccharides.

435. Novo's plant-scale design recommendations (Hemmingsen, 1979), based on the criteria for their Sweetzyme^® such as initial enzyme activity 200 IU/gm, half-life 825 hr, running time 2 half-lives, and wet bulk density 0.3 gm/cm³, are as follows:

As Novo indicates, similar data have been employed as a basis for the design and engineering of a number of fructose syrup plants in Europe, the Far East, the United States, Canada, and Latin America (Hemmingsen, 1979).

436. In 1977-82 Cetus Corporation, the U.S. R&D company headquartered in Berkeley, California, developed a patented two-step process for making 100% pure fructose from glucose syrups (U.S. Patents 4,247,641; 4,284,723), which is "ready for scale-up" (McGraw-Hill's Biotechnology Newswatch, 1982a). The project had funding of some $8 million from the Standard Oil Company of California (Socal), which, having retained its 17% corporate investment in Cetus, had left the latter with full title to all patents and know-how generated. Cetus' proprietary new technology converts the glucose to 100% fructose in two steps - one enzymatic (the oxidation of glucose to D-gluconolactone by means of immobilized glucose oxidase derived from Polyporus obtusus), the other chemical (the reduction of the gluconolactone to fructose by hydrogen with palladium as a catalyst). The hydrogen peroxide that appears as a by-product of the enzymatic glucose oxidation is utilized to transform ethylene or propylene into the corresponding oxide by the consecutive action of the two other enzymes, haloperoxidase and halohydrin epoxidase (Katchalski-Katzir & Freeman, 1982; Borglum & Marshall, 1984). A Cetus representative termed the process developed for the transformation of glucose as "the third-generation technology for fructose production" (McGraw Hill's Biotechnology Newswatch, 1982a). Cetus expected to develop the process to a full-scale facility of 250,000 to 500,000 tons annual capacity (Borglum & Marshall, 1984). Apparently the method was never scaled up since the enzyme has not become available in large quantities (Jensen & Rugh, 1987), and Cetus seems to have abandoned the concept.

4.     Economic estimates

437. The increase in high fructose syrup penetration of the world sweetener market was made possible in part by the evolution of immobilized glucose isomerase technology and improvement in enzyme characteristics and process conditions. Each improvement enhanced cost effectiveness. Shown below, for example, are the dynamics of decreasing the relative enzyme cost for producing high fructose corn syrup by Novo (Carasik & Carroll, 1983).

438. Economic estimates are available for the Kyowa Hakko process of isomerization of glucose by means of glucose isomerase immobilized on phenol-formaldehyde resin (see paragraph 428), which can be compared with data obtained from the use of native (soluble) glucose isomerase (Yokote et al., 1975). The cost estimate is based on a process in which 50 tons of glucose are converted monthly to an isomerized sugar mixture containing 45% fructose. In order to isomerize 1,000 kg of glucose, 21 kg of glucose isomerase containing fungal mycelia are required in the batch process with native enzyme. In the case of the immobilized enzyme systems 9.8 liters of immobilized enzyme can be obtained from the same amount of mycelia. Being adsorbed on Duolite A7, 9.8 liters of the immobilized enzyme can isomerize 2,822 kg of glucose within 30 days when operated at 60°C with 40% glucose solution as substrate. It is also presumed that an immobilized enzyme system is operated with two columns in series and that older columns are replaced with fresh columns every 30 days. MgSO₄ and CaCl₂ requirements in the immobilized glucose isomerase systems are 10% and 5% respectively of the necessary amounts of MgSO₄ and CaCl₂ with the native enzyme systems, thereby decreasing the volume of ion exchange resins required for deionization of the isomerized sugar mixture. Coloration of the isomerized sugar mixture in the immobilized enzyme system is much smaller, i.e. about 1/100, compared with that of the native enzyme system, thus also decreasing the cost for decolorization. Labor costs are also reduced in the immobilized enzyme system compared with the batch system. Taking into account all of these factors, the relative projected costs of these two systems are calculated (Figure 6). The cost of the Duolite A7 system is 61.5% of the cost of the batch system.

439. Purdue University presented another set of economic estimates (Emery et al., 1976) concerning an investigation of the optimum cycle time and relative costs of operation for soluble and immobilized glucose isomerase in which the latter is covalently bonded "to an expensive but good carrier", i.e., agarose activated by cyanogen bromide. The goals of the investigation were to bind the enzyme to the carrier, measure activity and stability of the preparation, and design a plant to process 450 tons/year of dry HFCS. Optimum cycle time is calculated to give results at minimum annual cost of the reaction. One-year carrier life is assumed. A batch process (soluble enzyme) is also designed for comparison purposes. For the batch process, the enzyme is the major cost, as shown below, while for the immobilized-enzyme process, the expensive carrier and reagents are the major costs. Even though these latter costs are high, the immobilized enzyme process is less costly because the expensive enzyme is used in the continuous process many times more frequently than in the batch. It is noteworthy that the actual isomerization step, including cost of the enzyme, usually accounts for less than a quarter of the total production cost, with pretreatment of feedstock and product clean up accounting for the remainder.