660
Bioconjugate Chem. 1994, 5, 660-665
Preparation and Characterization of a Bifunctional Fusion Enzyme Composed of UDP-galactose 4-Epimerase and Galactose-1-P Uridylyltransferase Yasushi Tamada,+,*Barbara A. Swanson,' Abolfazl Arabshahi,' a n d Perry A. Frey*,+ Institute for Enzyme Research, The Graduate School, and Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisconsin, and Tsukuba Research Laboratory, J a p a n Synthetic Rubber, 25 Miyukigaoka Tsukuba, Ibaraki 305, Japan. Received March 8, 1994@
A fusion enzyme consisting of UDP-galactose 4-epimerase and galactose-1-P uridylyltransferase with a n intervening Alas linker was constructed by in-frame fusion of E. coli gene galT to the 3'-terminus of the E. coli gene galE that had been extended with the coding sequence for three alanine residues, all contained within a high-expression plasmid. The fusion enzyme was expressed in E. coli and purified 24-fold to about 98% homogeneity by chromatography on hydroxylapatite and Q-Sepharose. On the basis of the comparison of the elution profile for enzyme activities upon gel permeation chromatography (Sephacryl S-400) with the molecular weight of 80 000 determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis, the fusion enzyme appears to exist in monomeric, dimeric, and tetrameric forms, all of which exhibit both enzymatic activities. The K, values of the fusion enzyme for substrates were similar to those for the corresponding native enzymes, except for UDPglucose, but the Kcat values were smaller than those for the native enzymes. The fusion enzyme shows kinetic advantages in that the initial velocity to produce glucose-1-P from UDP-galactose and galactose1-P is about 20% faster than that for a mixture of equal activities of the separate enzymes.
Commercial and medical applications of enzymes are often enhanced by immobilization to a solid support (13). In some applications, two or more enzymes acting sequentially in a process may be coimmobilized. In cases such as the latter, there may be advantages to using fusion proteins formed from enzymes acting in sequence. A potential advantage includes the recycling of a n expensive cofactor, which may be covalently attached to the fusion protein through a flexible spacer. Another is that the product of one enzyme, being a substrate for the next, might be captured by the second enzyme because of its proximity and gives a kinetic advantage. Techniques have been reported for obtaining these benefits in enzymatic production systems (4). In the most typical method, the enzymes are chemically crosslinked. However, crosslinking can be difficult to control and produces a multiplicity of species. Fusion enzymes resulting from in-frame fusion of the genes specifying the respective enzymes, as originally described by Bulow et al. (5), lead to specific products, the structures and composition of which can be controlled through manipulation of the structures of the fusion genes. UDP-galactose 4-epimerase and galactose-1-P uridylyltransferase, hereafter referred to as epimerase and transferase, are key enzymes in the Leloir pathway for galactose metabolism (6). Galactose in the cell is initially phosphorylated by ATP to galactose-1-P by the action of galactokinase. This is followed by the reaction of UDPglucose' and galactose-1-P to give UDP-galactose and glucose-1-P, catalyzed by galactose-1-P uridylyltrans-
* Address for correspondence: Institute for Enzyme Research, University of Wisconsin-Madison, 1710 University Ave., Madison, WI 53705. ' Institute for Enzyme Research, University of WisconsinMadison. Tsukuba Research Laboratory, Japan Synthetic Rubber, 25 Miyukigaoka Tsukuba, Ibaraki 305, Japan. Abstract published in Advance ACS Abstracts, September 15, 1994. @
ferase. In subsequent steps, phosphoglucomutase catalyzes the isomerization of glucose-1-P to glucose-6-P, and UDP-galactose 4-epimerase catalyzes the isomerization of UDP-galactose to UDP-glucose. UDP-galactose 4-epimerase7 from E. coli, is a dimer of identical subunits with a n overall molecular weight of 79 000 (7). It contains one tightly but noncovalently bound NAD+ or NADH per subunit. The crystal structure was determined by Bauer et al. (8). The catalytic mechanism includes the reversible oxidation of UDPgalactose or UDP-glucose to UDP-4-ketoglucose and concomitant reduction of NAD+ to NADH a t the active site. Galactose-1-P uridylyltransferase from E. coli consists of two identical subunits and has an overall molecular weight of 80 000 (9). The enzymatic mechanism is a double-displacement of the uridylyl group that obeys ping-pong kinetics and requires the formation of a covalent uridylyl-enzyme as the intermediate (10). In the intermediate, the uridylyl group is bonded to histidine 166 (11). We have begun a study of fused enzymes by creating a bifunctional fusion protein composed of UDP-galactose 4-epimerase and galactose-1-P uridylyltransferase. These two enzymes were chosen because they have been cloned, expressed, and characterized, and the product of one enzyme is a substrate for the other. This relationship could allow a fusion enzyme composed of them to catalyze the interconversion of galactose-1-P and glucose-1-P without the addition or intermediate formation of free nucleotide sugars, which are expensive and labile, provided that the nucleotide sugars can be covalently attached. In Scheme 1, the nucleotide sugars are shown Abbreviations: UDP-glucose, uridine 5'-diphosphoglucose; UDP-galactose, uridine 5'-diphosphogalactose; NAD+, nicotinamide adenine dinucleotide; NADP', nicotinamide adenine dinucleotide phosphate; NADH, reduced NAD+; NADPH, reduced NADP+; PMSF, phenylmethanesulfonyl fluoride; EDTA, ethylenediaminetetraacetic acid; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis.
1043-180219412905-0660$04.50/0 0 1994 American Chemical Society
'
Bioconjugate Chem., Vol. 5, No. 6,1994 661
Preparation of a Bifunctional Fusion Enzyme
the chain-termination method using Sequenase version 2.0 according to the protocol. A detailed description of the construction of the fusion gene, pT7ET, is given in the Results and Discussion.
Scheme 1
r\
UDPGlc GalE LGalT
UDPGa12GalT
Gal-1-P
Glc-1-P
attached through a long spacer to the fusion linkage between the enzymatic components, but in practice the spacer could be attached to either component, provided it is long enough. The spacer must be long, flexible, and inert under the conditions of the enzymatic reaction. In this paper, we describe the construction of a fusion gene between gaZE and gaZT. We also describe the expression, purification, characterization, and some of the properties of this fusion protein. EXPERIMENTAL PROCEDURES
Chemicals and Reagents. All restriction enzymes and polynucleotide kinase were purchased from New England Biolabs, and T4 DNA ligase and alkaline phosphatase were purchased from Boehringer Mannheim. UDP-glucose, UDP-galactose, UDP-glucose dehydrogenase, galactose-6-P dehydrogenase, phosphoglucomutase, NAD+, NADP+, ampicillin, streptomycin sulfate, and PMSF were purchased from Sigma. Ammonium sulfate (enzyme grade) was purchased from SchwarziMann Biotech. Bacto tryptone and bacto yeast extract were purchased from Difco Labs. The reagents and enzymes for DNA sequencing were from the Sequenase version 2.0 kit (United States Biochemical), and [a-35SldATP(600 Ci/mmol) was purchased from Amersham. Hydroxylapatite (Bio-Gel HTP) was the product of BioRad, and Q-Sepharose (fast flow) was purchased from Sigma. Sephacryl S-400 high resolution was the product of Pharmacia. The reagents for oligonucleotide synthesis were obtained from Glen Research Co. All other chemicals were obtained in reagent grade from commercial suppliers. Bacterial Strains and Plasmids. E. coli JMllO (dam, dcm, sup E44hsdR17phi leu rpsL lacy galK galT ara t o d thr tsx D(lac-proAB)F[traD36 proAB+ lac14 lacZ DM151) was purchased from Strategy Gene, DH5a (supE44 DlacU169 (BOlacZ DM15) hsdR17 recAl endAl gyrA96 thi-l relAl) from Gibco BRL Life Technology, and BL21(DE3) (hsdS gal (xlIts857 indl sam7 nin5 lacUVT7 genel)) pLysS was from Novagen, Inc. Because the BL21 cell strain does not contain the gal operon, it does not produce any endogenous epimerase or transferase. The LysS plasmid encode for a small lysozyme and the presence of this enzyme tends to decrease the loss of construct plasmid during cell growth (11). Plasmid pT7E2 containing the gaZE gene was constructed as described by Swanson and Frey (121, and the plasmid pKFt containing the galT gene was described by Kim et al. (13). Construction of Plasmid. Restriction enzyme digestion and other enzymatic reactions for DNA manipulations were performed as directed by the manufacturers. Oligonucleotides to make the linker and a primer for DNA sequencing were synthesized using a Biosearch Model 8600 DNA synthesizer and purified according the procedure as described (14). The purification of DNA was performed using QIAGEN columns according to the specified procedure. DNA sequencing was performed by
Expression and Purification of Fusion Enzyme. E. coli BL21(DE3), pLysS, pT7ET cells were streaked out on LB plates and grown a t 37 "C overnight. A single colony was used to inoculate 5 mL of 2xYT medium with ampicillin (0.1 mg/mL) and chloramphenicol (0.1 mg/mL), which was shaken a t 37 "C for 16 h. This starter culture was used to inoculate 1 L of 2xYT medium containing ampicillin (0.1 mg/mL). The growth medium was shaken a t 37 "C for 9 h with the addition of ampicillin (0.1 mg/ mL) every 3 h. The cells were harvested by centrifugation, quick-frozen in liquid nitrogen, and stored a t -70 "C .
The purification scheme for the fusion enzyme is similar to that for the purification of epimerase (15). All procedures were carried out a t 0-5 "C, and all buffers contained 1 mM EDTA, 10 mM 2-mercaptoethanol, and 1 mM PMSF. All centrifugations were a t 14000g. The cells from a 1L growth were thawed and suspended in 5 mL of 20 mM potassium phosphate buffer ( K P i buffer) a t pH 7.4. The pLysS plasmid encodes a lysozyme which effectively served to lyse the cells within 40 min. Trace amounts of deoxyribonuclease (DNase) and CaClz were added to the solution. After cell debris was removed by centrifugation, streptomycin sulfate (10%) was slowly added to the supernatant to a concentration of 25.4 mg/ mL. The solution was stirred for a n additional 20 min and centrifuged to remove the precipitate. Finely ground ammonium sulfate was slowly added to the supernatant fluid with stirring to 45% of saturation. The suspension was stirred for an additional 30 min and then centrifuged. The pellet was resuspended in a minimum of 20 mM KPi buffer a t pH 7.4 and dialyzed against the same buffer overnight. This crude protein solution (ca. 5 mL) was loaded onto a 20 mL column of hydroxylapatite that had been equilibrated with 20 mM KPi buffer a t pH 7.4. The column was eluted with the same buffer, and fractions containing both epimerase and transferase activities were pooled and carefully titrated to pH 8.5 with KOH. This protein solution was then loaded onto a 20 mL Q Sepharose column t h a t had been equilibrated with 20 mM K p i buffer a t pH 8.5. The protein was eluted by a linear gradient of KF'i buffer 200 mL in total volume and increasing from 20 mM to 300 mM a t pH 8.5. Fractions containing both enzyme activities were pooled and concentrated to 1 mL in a Centriprep-30 concentrator (Amicon). This protein solution was subjected to gel filtration on a Sephacryl S-400 high resolution column (1 x 90 cm) t h a t had been equilibrated with 20 mM KPi buffer containing 100 mM KC1 a t pH 7.4. Fractions from the main peak showing both enzyme activities were pooled and concentrated by a Microcon 10 concentrator (Amicon). Enzyme Assay. Galactose-1-P uridylyltransferase activity was assayed in a coupled assay (16) in which the formation of glucose-1-Pis coupled to NADPH production. The reaction mixture consisted of 0.2 mM galactose-1-P, 0.051 mM UDP-glucose, 0.326 mM NADP+, 3.3 pM glucose-l-6-Pz, and excess phosphoglucomutase and glucose-6-P dehydrogenase in 0.1 M sodium bicinate a t pH 8.5 a t 27 "C. The reaction was started by adding the test enzyme and monitoring NADPH formation a t 340 nm. The activity unit was defined to be the production of 1 pmol of glucose-1-P per hour. Epimerase activity was measured using the coupled assay described by Wilson and Hogness (17), in which the formation of UDP-glucose
662 Bioconjugafe Cbem., Vol. 5, No. 6,1994
Tamada et al.
linker galE
-
-
Ala Ala Ala
I--
5'- a C C G A T g CGGCCGCGATGACGCAATITAATCCCG-
EcoRV Linker
NotI
1
EcoRV + prll
GATCATCTAGAGlTAACGGCGCCAGCTAGCTAGTGCA-3' BcII Narl Pstl Figure 2. DNA sequence of the linker used to fuse the genes
galE and galT. Shown is the coding strand for the synthetic double-standed linker used in the fusion of the genes for UDPgalactose 4-epimerase and galactose-1-P uridylyltransferase in Figure 1.
J.
galE +
pT7ET
Figure 1. Construction of the plasmid pT7ET encoding an inframe fusion gene between galE and galT.
is coupled to NADH formation by UDP-glucose dehydrogenase. The enzyme was added to UDP-galactose (0.05 mM), NAD+ (0.25 mM), UDP-glucose dehydrogenase (200 units), and 0.125 M potassium bicinate buffer a t pH 8.5 at 27 "C; NADH formation was monitored a t 340 nm. One unit of enzyme is defined a s the amount catalyzing the formation of 1 pmol of UDP-glucose per hour. To determine the catalytic activity of the fusion enzyme, the initial rate was measured by a modification of standard assay for the transferase, in which UDP-galactose was substituted for UDP-glucose. The fusion enzyme was added to UDP-galactose (0.051 mM), NADP+ (0.326 mM), glucose-1,6-Pz (3.3 pM), and excess phosphoglucomutase and glucose-6-P dehydrogenase in 0.1 M sodium bicinate buffer a t pH 8.5 and varying amounts of galactose-1-P a t 27 "C; NADPH formation was monitored a t 340 nm. Electrophoresis. SDS-PAGE was performed using a Phast Gel System (Pharmacia) on a 8-25% linear gradient polyacrylamide gel. The protein bands were visualized by staining with Coomassie brilliant blue G250. The molecular weight of the enzyme subunits was determined by comparing their relative mobilities with those of proteins in a molecular weight standard kit (Pharmacia). RESULTS AND DISCUSSION Construction of pT7ET. The strategy used for the construction of a plasmid (pT7ET) containing a fusion gene composed ofgalE andgalT separated by an in-frame linker coding for Ala3 is illustrated in Figure 1. A 75 bp DNA fragment, whose sequence in the coding strand is shown in Figure 2, was prepared by synthesis of the complementary single strands and hybridization. It contains a coding sequence for the 3'-end of galE that eliminates the stop codon and then encodes a linker of three alanine residues. The last part of the oligomer is the 5'-end of the galT start codon and the BclI site. The
unique NotI restriction site was designed into the linker region of the fragment sequence in order to facilitate a check of the construction of the fusion gene after cloning and to allow for future addition of linkers into this site. This oligomer sequence contains BclI and NurI restriction sites to facilitate insertion of the galT gene. As shown in Figure 1,the linker fragment was ligated into the PstI and EcoRV sites of pT7E2. The plasmid obtained, pT7E(75), was transformed into DH5a cells. Purified plasmid was linearized by cleavage with NarI, dephosphorylated, and then digested with BclI. This procedure efficiently prevented self-ligation. The galT gene was prepared by digesting p u t , grown in JMllO cells to prevent dam methylation, with NarI and BclI. Because pKFt has five NarI sites and two BclI sites, eight fragments were obtained from pKFt. The 1095 bp fragment, containing most of the galT gene, was isolated and purified, using low melting temperature agarose gel electrophoresis, and then ligated into the linearized pT7E(75). The plasmid obtained, pT7ET, encoding for a fusion enzyme, was checked by restriction analysis and by sequencing the linker region in the fusion gene. The pT7ET plasmid confers ampicillin resistance and contains the T7 promoter just ahead of the fusion gene. The fusion gene in pT7ET was efficiently expressed, as determined by transformation into E. coli BL21 and SDS-PAGE analysis of cell extracts. The fusion protein appeared as a prominent band corresponding to a molecular weight of 80 000, which was absent in untransformed cells (see, for example, lane 1 of Figure 3). Purification of Fusion Enzyme. Purification of the fusion enzyme was initially attempted according to protocols developed for the epimerase and transferase. The fusion enzyme did not bind to the Mi-Gel Blue column used in the transferase procedure; however, applicaton of the procedure for purifying the epimerase led to the purified fusion protein in a high degree of homogeneity. The results are summarized in Table 1, and in each step the purity of the fusion enzyme was confirmed by SDS-PAGE as shown in Figure 3. On the basis of the results of SDS-PAGE, the purified fusion enzyme was estimated to be approximately 95% homogeneous. In addition to the original epimerase protocol, gel filtration on Sephacryl S-400 high resolution was employed as the final step of the purification. The elution profile after gel filtration shown in Figure 4 shows that the two enzyme activities are copurified. Selected fractions corresponding to the major protein peak in Figure 4 were pooled as the main form of the fusion protein. The purified fusion enzyme exhibits 2460 units of epimerase activity per mg of protein and a transferase specific activity of 1000 U/mg. These specific activities are less than those of the corresponding native enzymes, which are normally 8000-10 000 U/mg for epimerase and 10 000-11 000 U/mg for the transferase in our laboratory. Therefore, the fusion protein exhibited about 25%
Bioconjugate Chem., Vol. 5, No. 6, 1994 663
Preparation of a Bifunctional Fusion Enzyme Table 1. Purification of Fusion Protein
ammonium sulfate hydroxylapatite Q-Sepharose Sephacryl S400
protein (mg) 320 27 4.8 2.7
epimerase activity total (units) specific (unitdmg) 108 34 600 27 200 1011 8500 1720 6470 2460
yield (%I 100 8.4 1.6 0.84
transferase activity total (units) specific (unitdmg) 37 400 117 543 14 600 3930 795 2910 1080
Table 2. Kinetics parameters of Fusion proteina
UDP-
-
+94,000 +67,000 +43,000 +30,000 20,100 14,400
1 2 3 4 5 Figure 3. SDS-PAGE of the fusion enzyme. Shown are 8%25% electrophoretic gels of the fusion enzyme at various points in its purification. The bands were stained with Coomassie Brilliant Blue G250. Lanes: 1, after ammonium sulfate precipitation; 2, after hydroxylapatite chromatography; 3, after Q-Sepharose chromatography; 4, after Sephacryl S-400 HR chromatography; 5,standards phosphorylase b (94 000), bovine serum albumin (67 OOO), ovalbumin (43 OOO), carbonic anhydrase (30OOO), soybean trypsin inhibitor (20 1001, a-lactalbumin (14 400).
o*2x8w
3
0
B
0
0.1
0.0 -.100
110
120
130
140
Fraction No.
Figure 4. Gel filtration chromatography of the fusion enzyme on Sephacryl S-400. The fusion enzyme was chromatographed through a 1 x 90 cm column of Sephacryl S-400 high resolution equilibrated and eluted with 20 mM KPi, 1mM EDTA, 10 mM 2-mercaptoethanol, 100 mM KCl, pH 7.4 at 4 "C: (0)OD 280, (A) UDP-galactose 4-epimerase activity, (A) galactose-1-P uridylyltransferase activity.
and lo%, respectively, of the activities of the individual enzymes. Part of the reason for the lower activities is t h a t the molecular weight of the fusion protein is twice that of either enzyme, so that the measured specific activity for each enzyme would be half that for the separate enzyme. On this basis the specific activities of the fusion enzyme are 50% and 20% of the activities of epimerase and transferase, respectively. On the basis of the activity of the crude extract from cells carrying the overexpression vector, the epimerase activity was purified 24-fold, and the transferase activity was purified 9-fold. Galactose-1-P uridylyltransferase is reported to be labile to proteolysis and oxidation (16). Because the transferase moiety in the fusion enzyme may
galactose 0.16
galactose1-P
UDP-
glucose
kat (s-l) 500
epimeraseb transferaseC 0.30 f 0.03 0.20 f 0.02 960 fusion protein epimerase 0.14 f 0.02 160 f 20 transferase 0.29 f 0.01 0.08 f 0.01 24 f 3 "K, and k,t values were determined by the initial rate measurements at six different concentrations according to the method described under enzyme assay. Wilson, D. B., and Hogness, D. S. (1964)J.Biol. Chem. 230,2469. Wong, L. J., and Frey, P. A. (1974)Biochemistry 13, 3889. also be unstable, its activity may have declined in the course of purification despite the addition of PMSF, EDTA, and 2-mercaptoethanol to the buffers. Molecular Weight of Fusion Enzyme. The molecular weight of the fusion enzyme was determined by gel filtration and SDS-PAGE. As shown in Figure 3, the subunit molecular weight of the fusion enzyme is estimated to be about 80 000 based on the results of SDSPAGE. This value is in good agreement with the expected size. Because UDP-galactose 4-epimerase and galactose-1-P uridylyltransferase are dimeric enzymes with total molecular weights of 79 000 and 80 000, respectively, the molecular weight of a subunit of the fusion enzyme should be about 80 000. Gel permeation chromatography of the fusion enzyme purified through the Sepharose Q step and comparison of the elution volumes with protein standards showed the fusion enzyme to be heterogeneous with respect to molecular weight. As shown in Figure 4, three peaks for the fusion enzyme were observed in the gel filtration profile. The peak at fractions 120-125 corresponds to a molecular weight of 60 000-80 000, the peak at fractions 112-117 corresponds to a molecular weight of 140 000160 000, and the third peak at fractions 107-110 is approximately 320 000-380 000 in molecular weight. These results indicate that the fusion enzyme exists in three forms, which appear to be monomeric (E-T), dimeric (E-T)z, and tetrameric (E-T)4. The monomeric fusion enzyme exhibits very low epimerase and significant transferase activity. The dimeric fusion enzyme exhibits both activities at high levels, and this is the peak containing the major total activity. The specific activities of both epimerase and transferase are higher in the apparent tetrameric form of the enzyme but could not be accurately estimated from the elution profile in Figure 4. Only a small percentage of the fusion enzyme exists in the tetrameric form, as indicated by the amount of protein associated with this form in the elution profile of Figure 4.
Kinetic Parameters for the Fusion Enzyme. Kinetic parameters for the fusion enzyme are summarized in Table 2, together with the corresponding values for the native epimerase and transferase determined under the same experimental conditions. The K , values of the fusion enzyme for galactose-1-P and UDP-galactose are similar to those of the corresponding native enzymes, but the K , value for UDP-glucose is much smaller than t h a t
Tamada et al.
664 Bioconjugate Chem., Vol. 5, No. 6,1994 0.02 I
I
C
E
E
si
a
[galactose-I-PI, mM
Figure 5. Initial rates of the coupled actions of fused epimerase-transferase and of the separate enzymes. Initial rates of the coupled actions of epimerase and transferase in the transformation of galactose-1-P into glucose-1-Pat various concentrations of galactose-1-P are plotted. The rates glucose-1-P formation were measured in assay for galadose-l-P uridylyltransferase assay method, except that UDP-galactose was substituted for UDP-glucose. The epimerase and transferase activities associated with a sample of the fusion enzyme (16.8 units and 1.0 unit, respectively) were matched by mixing the same amounts of the separate enzymes assayed at the same time. Symbols: ( 0 )fusion enzyme, (0) mixed native epimerase and transferase. of the transferase. During the kinetic determinations, UDP-galactose can be converted by the epimerase moiety into UDP-glucose, effectively increasing the concentration of this compound from what was initially added. This effect would result in an apparently smaller K, value of the fusion protein for UDP-glucose. The kcat values of each moiety in the fusion enzyme are somewhat smaller than those of the corresponding native enzymes. The wild type transferase is more labile than the epimerase, owing to its greater susceptibility to oxidation and cleavage by adventitious proteases. This is also true of the fusion enzyme. The activity of the transferase component decreases with time more rapidly than t h a t of the epimerase component. The wild type transferase is stabilized by cysteine or dithiothreitol, the absence of dioxygen, and the presence of proteolytic inhibitors; it seems likely that the transferase component of the fusion enzyme can be similarly stabilized. Proximity Effect. The fusion enzyme contains both epimerase and transferase activities and so will catalyze the transformation of galactose-1-P into glucose-1-P in the presence of a catalytic amount of UDP-galactose (or UDP-glucose) according to eq 3. This results from the coupled actions of the epimerase (eq 1)and transferase (eq 2). The rate of the conversion of galactose-1-P into UDP-galactose == UDP-glucose UDP-glucose
+ galactose-1-P ==
UDP-galactose
(1)
+ glucose-1-P (2)
sum: galactose-1-P * glucose-1-P
(3)
glucose-1-P in the presence of UDP-galactose was measured in a modification of the transferase assay. The results are shown in Figure 5, together with the results obtained from a matched mixture of the two native enzymes. The epimerase and transferase activities of the fusion enzyme used in Figure 5 were 16.8 units and 1 unit, respectively. In comparing the rates produced by the fusion enzyme with those produced by the mixed native enzymes, the same amounts of each native en-
zyme, 1unit of transferase and 16.8 units of epimerase, were used. As shown in Figure 5, the initial rates generated by the fusion enzyme are faster than those from the mixed native enzymes at all substrate concentrations used in this experiment. This result suggests t h a t a proximity effect is operating in the fusion enzyme to increase the catalytic activity for the overall reaction. Inasmuch as the distance between the UDP-galactose 4-epimerase moiety and the galactose-1-P uridylyyltransferase moiety in the fusion enzyme could be much less than t h a t between the two unfused enzymes, UDPglucose produced by the epimerase might be captured by the transferase faster in the fusion enzyme than between the two separate enzymes. This form of substrate channeling effect improves the overall performance of the fusion enzyme (18). While the channeling in Figure 5 is significant, it is not a great or important effect. The Enzyme Linker. The trialanyl linker connecting the epimerase and transferase in the fusion enzyme may influence its physical conformation and enzymatic activities. In this study a short linker, three alanine residues, was chosen because it had been reported that longer linkers are susceptible to proteolytic degradation during expression, resulting in a low yield of the fusion enzyme (19). However, three alanyl residues may not provide sufficient flexibility to allow for optimal folding and subunit-subunit association of both enzymes in the fusion enzyme. This could explain why the catalytic activities for each moiety of the fusion enzyme were only 50%and 20% of those of the native epimerase and transferase, respectively. It is also possible that reversing the sequence of the two genes in the construct could affect channeling efficiency. In the future, other linkers will be introduced into the fusion enzyme in order to investigate the role of the linker region and perhaps to prepare a new fusion enzyme exhibiting higher, more stable, or more efficiently coupled enzymatic activities. The linker region in the fusion gene is designed to allow extensions of the linker region in the fusion protein through the introduction of synthetic double-stranded oligodeoxynucleotides into the unique Not1 site. ACKNOWLEDGMENT This research was supported by Grant No. GM 30480 from the National Institute of General Medical Sciences of the U.S. Public Health Service. LITERATURE CITED (1) Chibata, I., Tosa, T., and Sato, T. (1991) Industrial applications of immobilized proteins. Protein Immobilization: Fundamentals and Applications (R.F. Taylor, Ed.) pp 339-362,
Marcel Dekker, Inc., New York.
(2) Chang, T. M. S. (1991) Therapeutic applications of immobilized proteins and cells. Protein Immobilization: Fundamentals and Applications (R.F. Taylor, Ed.) pp 305-318,
Marcel Dekker, Inc., New York. (3) Scouten, W. H. (1981)A,@nity Chromatography Bioselectiue Adsorption on Inert Matrices, pp 305-337, John Wiley 8z Sons, New York. (4) Cabral, J. M. S., and Kennedy, J. F. (1991) Covalent and coordination immobilization of proteins. Protein Immobilization: Fundamentals and Applications (R.F. Taylor, Ed.) pp 73-138, Marcel Dekker, Inc., New York. (5) Bulow, L. (1987) Characterization of an artificial bifunctional enzyme, P-galactosidaselgalactokinase,prepared by gene fusion. Eur. J.Biochem. 163, 443-448. (6) Caputto, R., Leloir, L. F., Trucco, R. E., Cardini, C. E., and Paladini, A. C. (1949) The enzymatic transformation of
Bioconjugate Chem., Vol. 5, No. 6,1994 665
Preparation of a Bifunctional Fusion Enzyme
galactose into glucose derivatives. J.Biol. Chem. 179, 497498. (7) Frey, P. A. (1987) Complex pyridine nucleotide-dependent transformations. Pyridine Nucleotide Coenzymes: Chemical, Biocemical, and Medical Aspect (D. Dolphin, R. Poulson, and 0. Avramovic, Eds.) pp 461-511, Wiley, New York. (8) Bauer, A. J., Rayment, I., Frey, P. A., and Holden, H. M. (1992)The molecular structure of UDP- alactose 4-epimerase from Escherichia coli determined at 2.5 resolution. Proteins 12, 372-381. (9) Saito, S., Ozutsumi, M., and Kurahashi, K. (1967) Galactose 1-phosphate uridylyltransferase of Escherichia coli: I1 Further Purification and characterization. J . Biol. Chem. 243, 2362-2368. (10)Frey, P. A., Wong, L. J.,Sheu, K. F., andYang, S. L. (1982) Galactose-1-phosphate uridylyltransferase: detection, isolation, and characterization of uridyly enzyme. Methods Enzymol. 87, 20-36. (11) Moffat, B. A., and Studier, F. W. (1987)Use of T7 lysozyme to improve the T7 gene expression system. Cell 49,221-227. (12) Swanson, B. A,, and Frey, P. A. (1993)Identification of Lys 153 as a functionally important residue on UDP-galactose 4-epimerase from Escherichia coli. Biochemistry 32, 1323113236. (13) Kim, J., Ruzicka, F., and Frey, P. A. (1990) Remodeling
x
hexose-1-phosphateuridylyltransferase: Mechanism-inspired mutation into a new enzyme, UDP-hexose synthase. Biochemistry 29, 10590-10593. (14) Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A laboratory manual, Vol. 2, Cold Spring Harbor Laboratory Press, New York. (15) Bauer, A. J., Rayment, I., Frey, P. A., and Holden, H. M. (1991) The isolation, purification, and preliminary crystallographiccharacterization of UDP-galactose 4-epimerase from Escherichia coli. Proteins 9, 135-142. (16) Wong, L. J., and Frey, P. A. (1974) Galactose-1-phosphate uridylyltransferase: Rate studies confirming a uridylylenzyme intermediate on the catalytic pathway. Biochemistry 13, 3889-3894. (17) Wilson, D. B., and Hogness, D. S. (1964) The enzymes of the galactose operon in Escherichia coli: I. Purification and characterization of uridine diphosphogalactose 4-epimerase. J. Biol. Chem. 239, 2469-2481. (18) Ljungcranyz, P., Carlsson, H., Mansson, M. O., Buckel, P., Moshach, K., and Bulow, L. (1989) Construction of an artificial enzyme, /I-galactosidasdgalactose dehydrogenase, exhibiting efficient galactose channeling. Biochemistry 28, 8786-8792. (19) Bulow, L., and Moshach, K. (1991) Multienzyme systems obtained by gene fusion. Trends Biothecnol. 9, 226-231.
