S-peptide as a potent peptidyl linker for protein cross-linking by microbial transglutaminase from Streptomyces mobaraensis

Bioconjugate Chem. 2003, 14, 351−357

351

S-Peptide as a Potent Peptidyl Linker for Protein Cross-Linking by Microbial Transglutaminase from Streptomyces mobaraensis Noriho Kamiya,† Tsutomu Tanaka,† Tsutomu Suzuki,§ Takeshi Takazawa,† Shuji Takeda,† Kimitsuna Watanabe,§ and Teruyuki Nagamune*,† Department of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan and Department of Integrated Biosciences, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa-shi, Chiba 277-8562, Japan. Received September 21, 2002; Revised Manuscript Received January 15, 2003

We have found that ribonuclease S-peptide can work as a novel peptidyl substrate in protein crosslinking reactions catalyzed by microbial transglutaminase (MTG) from Streptomyces mobaraensis. Enhanced green fluorescent protein tethered to S-peptide at its N-terminus (S-tag-EGFP) appeared to be efficiently cross-linked by MTG. As wild-type EGFP was not susceptible to cross-linking, the S-peptide moiety is likely to be responsible for the cross-linking. A site-directed mutation study assigned Gln15 in the S-peptide sequence as the sole acyl donor. Mass spectrometric analysis showed that two Lys residues (Lys5 and Lys11) in the S-peptide sequence functioned as acyl acceptors. We also succeeded in direct monitoring of the cross-linking process by virtue of fluorescence resonance energy transfer (FRET) between S-tag-EGFP and its blue fluorescent color variant (S-tag-EBFP). The protein cross-linking was tunable by either engineering S-peptide sequence or capping the S-peptide moiety with S-protein, the partner protein of S-peptide for the formation of ribonuclease A. The latter indicates that S-protein can be used as a specific inhibitor of S-peptide-directed protein cross-linking by MTG. The controllable protein cross-linking of S-peptide as a potent substrate of MTG will shed new light on biomolecule conjugation.

INTRODUCTION

The major choice for conjugation or covalent modification of functional proteins with an organic molecule (including peptides and proteins) has basically been chemical manipulation because of their simplicity and variety (1). Although choosing the correct reagent systems can afford successful modification, one of the critical shortcomings in the chemical modification of proteins is the lack of specificity because of a number of functional groups on the protein surface that compete for a target molecule. Genetic manipulation, in which gene fragments encoding different kinds of peptides and/or proteins are fused onto a plasmid vector, has also been widely employed to prepare recombinant fusion proteins. However, this approach is often hampered by the poor yield of the target chimeric proteins, mainly due to the formation of inclusion bodies in the host cells of foreign origins. Transglutaminase (TG), which strictly catalyzes the acyl-transfer reaction between the γ-carboxyamide group of a Gln residue (acyl donor) and the -amino group of a Lys residue (acyl acceptor) of proteins and peptides, is conceptually applicable to Gln- or Lys-specific covalent modification or cross-linking of target proteins via an -(γglutamyl)lysine bridge (2, 3). However, this concept was demonstrated to only be applicable to a limited number of proteins by the use of guinea pig TG due to its strict substrate specificity (4, 5). * To whom correspondence should be addressed. Tel: +81-3-5841-7328; Fax: +81-3-5841-8657; E-mail: nagamune@ bio.t.u-tokyo.ac.jp. † Department of Chemistry and Biotechnology. § Department of Integrated Biosciences.

Microbial transglutaminase (MTG) from Streptomyces mobaraensis (6, 7) was found to be a powerful industrial catalyst owing to the rather broader substrate specificity for acyl acceptors than other TGs (8), the handling feasibility, such as Ca2+-independence, and the availability originating from mass production by the microorganism. In the past decade, MTG has been an ideal enzyme for food processing (9), but there have been few reports on the application of MTG in other fields (10). Recently, the utilization of MTG has attracted much attention for either the modification or conjugation of functional proteins. Simple incorporation of organic molecules into target proteins was demonstrated in the preparation of a hapten-protein conjugate (11), a biotinylated monoclonal antibody (12), and a poly(ethyleneglycol)-modified human interleukin 2 (8). The conjugation of functional proteins (13) yielded a bifunctional protein. Site-specific modification of a peptide (14) or a protein (15) was directed to obtain an artificial nutrient. The stabilization of enzymes by MTG has also been studied by immobilization (16) and the elimination of deleterious factors (17). These studies clearly showed that MTG can be applicable to the modification of protein function and production of a new protein conjugate. However, most studies have been carried out with intact proteins (1013) and little information is yet available on the substrate specificity of MTG (6, 18-21). As claimed by Sato et al. (5), the capability of posttranslational modification by TG offers a new choice for site-specific modification of target proteins when a specific peptide sequence can be introduced to proteins of interest by genetic engineering. More attention should thus be paid to the substrate requirements of MTG, which is superior to other TGs from a

10.1021/bc025610y CCC: $25.00 © 2003 American Chemical Society Published on Web 02/20/2003

352 Bioconjugate Chem., Vol. 14, No. 2, 2003

practical standpoint, to generalize TG-directed enzymatic protein conjugation in the field of protein engineering. The aim of this study was to demonstrate that ribonuclease S-peptide can be a novel candidate as a tunable peptide tag that enables the site-specific cross-linking of functional proteins. Enhanced green fluorescent protein carrying an additional S-peptide sequence in its Nterminus (S-tag-EGFP) was employed as a model protein. The reactive cross-linking sites of dimerized S-tag-EGFP were explored by either site-directed mutagenesis or mass spectrometric analysis of S-tag-EGFP. Direct spectroscopic measurement of the cross-linking reaction was successfully monitored by the use of fluorescence resonance energy transfer (FRET) between cross-linked Stag-EGFP and S-tag-EBFP (a fluorescent color variant of EGFP). Finally, as we succeeded in either the alteration of the reactivity of S-peptide by site-directed mutagenesis or the inhibition of the cross-linking by capping the S-peptide moiety with S-protein, this report shows, for the first time, controllable protein cross-linking by TG. MATERIALS AND METHODS

Materials. MTG was generously provided by Ajinomoto Co. Inc. (Japan) in a semipurified form. Carbobenzoxy-L-glutamylglycine (Z-QG) was obtained from Tokyo Chemical Industry Co. Ltd (Japan). S-Protein was purchased from Sigma-Aldrich. All other reagents used in this study were of the highest purity commercially available. Sample Preparation. Purification of MTG was conducted with a Blue Sepharose CL-6B affinity column (Amersham Pharmacia Biotech Co.) as reported previously (6). The catalytic activity of purified MTG was measured by a colorimetric procedure described elsewhere (22). S-tag-EGFP and S-tag-EBFP were prepared as follows. DNA fragments encoding EGFP (enhanced GFP) and EBFP (enhanced blue fluorescent protein) were prepared from plasmids pEGFP and pEBFP-N1 (Clontech), respectively, and subcloned into pET32b(+) (Novagen) with NcoI/NotI double digestion. Both proteins were expressed in Escherichia coli strain BL21(DE3)pLysS (Novagen) as E. coli thioredoxin fusion proteins according to a previously described procedure (23). The fusion proteins were purified by metal affinity chromatography through His-tag (Clontech). The purified proteins were subjected to site-specific cleavage with thrombin (Novagen) to eliminate the thioredoxin and His-tag in the N-terminus of the fusion proteins. The resultant proteins were employed as S-tag-EGFP and S-tag-EBFP, respectively. The N-terminal extended peptide sequence of S-tag-EG(B)FP in comparison with wild-type EG(B)FP was GSGMKETAAAKFERQHMDSPDLGTDDDDKA (the S-peptide sequence is shown in bold letters). Sitedirected mutagenesis of Gln15 to Ala of S-tag-EGFP was performed using QuikChange mutagenesis kit (Stratagene) and the resultant protein was abbreviated as Q15Atag-EGFP. The primer nucleotide sequences used for the mutagenesis of Gln15 to Ala were 5′-GCT GCT GCT AAA TTC GAA CGC GCG CAC ATG GAC AGC-3′ and 5′-GCT GTC CAT GTG CGC GCG TTC GAA TTT AGC AGC AGC-3′. Site-directed mutagenesis of Lys5 to Ala of S-tagEGFP, which yielded K5A-tag-EGFP, was also performed by the same procedure. The primer nucleotide sequences used for the mutagenesis of Lys5 to Ala were 5′-GAA ACC GCT GCT GCT GCA TTC GAA CGC CAG CAC-3′ and 5′-GTG CTG GCG TTC GAA TGC AGC AGC AGC GGT TTC-3′. The homogeneity of the purified proteins was

Kamiya et al.

verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The protein concentration was determined using the BCA assay Kit (Pierce). In the case of S-tag-EGFP and S-tag-EBFP, the protein concentrations were adjusted based on the absorbance of chromophores by assuming millimolar extinction coefficients of 55 mM-1 cm-1 at 488 nm and 31 mM-1 cm-1 at 380 nm, respectively (24). Cross-Linking of S-tag-EGFP by MTG. In preliminary experiments, wild-type EGFP or S-tag-EGFP (1.2 µM) was dissolved in 10 mM sodium phosphate buffer at pH 7.5. The cross-linking reaction was initiated by the addition of MTG (0.25 U/mL) at 4 °C. After incubation for 2 h, the reaction products were analyzed by SDSPAGE. Image analysis of the result of SDS-PAGE was performed on a Macintosh computer using the public domain NIH Image program (developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). Identification of Cross-Linking Sites of S-tagEGFP by Mass Spectroscopy. Z-QG was used as an acyl donor to identify the cross-linking sites of Lys residues working as acyl acceptors in S-tag-EGFP. Incorporation of Z-QG into S-tag-EGFP was carried out using the same protocol described above except for the presence of an excess amount of Z-QG (250 µM) in the reaction mixture. The complete suppression of dimerization of S-tag-EGFP molecules by Z-QG was confirmed by SDS-PAGE. The reaction products on SDS-PAGE were visualized by CBB staining. The protein spot was excised and the gel piece was soaked in a buffer containing 0.2 M NH4HCO3 with 50% acetonitrile and incubated at 30 °C for 30 min to remove SDS from the proteins. To ensure the complete removal of SDS, the step was repeated twice. The gel pieces were dried in a speedvac and then rehydrated with 5-10 µL of trypsin digestion solution [0.2 M NH4HCO3, 15 ng/µL TPCK-trypsin (Pierce)]. After the gel pieces had absorbed the liquid, 10-20 µL of 0.2 M NH4HCO3 was added to the gel. Complete digestion was carried out overnight at 30 °C. The digested peptides were extracted from the gel by shaking in 200 µL of 60% acetonitrile and 0.1% TFA solution for 20 min. This step was repeated twice more. The collected fractions containing each peptide were pooled together and dried in a speedvac, and the obtained sample was dissolved in 80 µL of 0.1% formic acid. Twenty microliters of the collected peptide was subjected to LC/MS analysis to obtain the peptide mass map. A Finnigan LCQ ion trap mass spectrometer (ThermoQuest) equipped with an electrospray ionization source was used for the peptide analysis. The LC/MS analysis was performed using an ODS reverse-phase column (Magic C18, 0.1 × 15 cm; Michrom BioResource) connected on-line to the electrospray interface. A solvent system consisting of 0.1% formic acid in H2O (A) and acetonitrile (B) was developed from 0% B to 70% B in 35 min at a flow rate of 50 mL/min using the Magic 2002 HPLC system (Michrom BioResource). The flow rate of sheath gas and the capillary temperature were kept at 55 arb and 235 °C, respectively. The zoom scan analysis and MS/MS experiment by collisioninduced dissociation using the data dependent scan (triple play) were performed in the range of 300-2000 m/z throughout the separation (25, 26). Monitoring the Cross-Linking Reaction of S-tagEGFP and S-tag-EBFP by FRET. When the MTGcatalyzed cross-linking process by FRET was monitored, S-tag-EBFP (1.2 µM) and S-tag-EGFP (1.2 µM) were mixed in a quartz cell at 15 °C. The cross-linking reaction was initiated by the addition of MTG (0.25 U/mL) to the

Controllable Protein Cross-Linking by Transglutaminase

Figure 1. (A) Extra N-terminal peptide sequence of S-tagEGFP. (B) SDS-PAGE analysis of the reaction products after MTG treatment of wild-type EGFP, S-tag-EGFP and Q15A-tagEGFP (lane 1, low molecular weight markers; lane 2, wild-type EGFP; lane 3, wild-type EGFP treated with MTG; lane 4, S-tagEGFP; lane 5, S-tag-EGFP treated with MTG; lane 6, Q15Atag-EGFP; lane 7, Q15A-tag-EGFP treated with MTG; lane 8, high molecular weight markers).

reaction mixture. The change in fluorescence spectra (excited at 380 nm) was recorded at a certain timeinterval after the addition of MTG using an F-2000 fluorescence spectrophotometer (Hitachi, Japan). The ratio of fluorescence emission intensity at 508 nm to that at 444 nm (I508/I444) was normalized by the value at time zero (i.e., without MTG) and used as the FRET signal. When the cross-linking was followed by SDS-PAGE, an aliquot was periodically withdrawn and immediately mixed with SDS-PAGE sample buffer (50 mM Tris, 2% SDS, 6% 2-mercaptoethanol) to terminate the crosslinking reaction. When investigating the effect of Sprotein on the cross-linking, S-protein (3.6 µM) was mixed with the solution containing both S-tag-EBFP and S-tagEGFP described above. After the mixture was incubated for 1 h on ice, the cross-linking reaction was initiated by the addition of MTG to the mixture at 15 °C. RESULTS AND DISCUSSION

Cross-Linking of S-tag-EGFP by MTG. The primal aim of this study was to search for a potent peptide linker that enabled site-specific cross-linking of functional proteins by MTG. During the preliminary exploration, we found that recombinantly expressed EGFP protein in E. coli was efficiently cross-linked by MTG. The protein is a fusion protein of EGFP, the N-terminus of which has thioredoxin, His-tag, a thrombin site and S-tag attached. Since wild-type EGFP was not susceptible to crosslinking by MTG (Figure 1B), similarly to guinea pig liver transglutaminase (27), the target peptide of MTG should be found in this fused peptide sequence. The removal of the N-terminal thioredoxin and His-tag portion by thrombin cleavage showed no effect on the cross-linking by MTG, indicating that the thioredoxin moiety did not participate in the transglutamination (data not shown). This fact prompted us to identify the extra N-terminal S-tag peptide of EGFP as a key peptide for protein crosslinking by MTG. Interestingly, selective oligomeric, but no polymeric, forms of S-tag-EGFP were observed after the treatment with MTG in addition to the residual band around the monomeric form (Figure 1B, lane 5). The oligomeric

Bioconjugate Chem., Vol. 14, No. 2, 2003 353

products may correspond to the dimer and the tetramer based on the molecular weight of S-tag-EGFP. The residual fraction around the monomeric form showed a slight shift in mobility in the SDS-PAGE compared to the intact monomer. This behavior may be ascribed to alteration of the charge distribution caused by the competitive hydrolysis of Gln to Glu (28) or the intramolecular cross-linking previously observed with guinea pig transglutaminase (27). It is worth noting that the trimer was not detectable by SDS-PAGE, suggesting poor reactivity between the monomer and dimer in the crosslinking reaction. Identification of Cross-Linking Sites in Oligomeric S-tag-EGFP. Since Gln15 is the sole glutamine residue in the S-tag sequence, we substituted this residue with Ala (Q15A-tag-EGFP) by site-directed mutagenesis to confirm Gln15 as an acyl donor for the transglutamination. As shown in Figure 1B (lane 7), the Q15A-tagEGFP remained intact even after the MTG treatment, indicating that Gln15 was the sole acyl donor in this system. With regard to the acyl acceptor for the transglutamination, there are three Lys residues (K5, K11, and K29) in the extra N-terminal sequence of S-tagEGFP (Figure 1A). S-tag-EGFP was conjugated with Z-QG as a small acyl donor by MTG to determine the reactive Lys residues for cross-linking. The product was then digested with trypsin and subjected to mass spectrometric analysis to identify the Z-QG incorporated peptides. The peptide mass map of Z-QG treated-S-tagEGFP was compared to that of nontreated S-tag-EGFP (Figure 2). We found the absence of specific peptides, FER and ETAAAK, both of which were derived from the Nterminal S-tag sequence in the Z-QG treated-S-tag-EGFP (Figure 2A). No difference was observed in the other detectable peptides, including the K29-containing peptide. In the place of these missing peptides, we could identify two Z-QG-incorporated peptides derived from the N-terminal S-tag sequence, ETAAAK(zQG)FER (Figure 2B) and GSGMK(zQG)ETAAAK (Figure 2C). Thus, K5 and K11 were determined to be the acyl acceptors for the cross-linking by MTG. We can conclude that the possible cross-linking sites in oligomeric S-tag-EGFP are Gln15Lys5 or Gln15-Lys11. Since relative intensity of GSGMK(zQG)ETAAAK was lower than that of ETAAAK(zQG)FER, it can be assumed that Lys5 is a minor site for the cross-linking reaction. However, we could not rule out other possible acyl acceptors, because Z-QG conjugation may not exactly reflect the protein cross-linking reaction. It should be noted that both the reactive Lys residues are located in the S-peptide sequence, indicating that MTG strictly recognizes a flexible region of the target protein. Direct Monitoring of the Protein Cross-Linking Process Using FRET. The fluorescent intensity of S-tag-EGFP appeared to be virtually identical to the original one even after the transglutamination [more than 95% of fluorescence remained intact (data not shown)]. These observation encouraged us to carry out real-time spectroscopic measurement of the MTG-catalyzed cross-linking reaction using FRET (Figure 3). The blue fluorescent color variant, S-tag-EBFP, was prepared for this purpose and the fluorescent spectral change was monitored in the presence of an equimolar amount of S-tag-EGFP. It was confirmed that wild-type EBFP was not cross-linked by MTG as well as wild-type EGFP (data not shown). As shown in Figure 4, the increase of fluorescent intensity at 508 nm was successfully monitored in the presence of MTG, while no

354 Bioconjugate Chem., Vol. 14, No. 2, 2003

Kamiya et al.

Figure 2. Liquid chromatography mass spectrometric analysis to identify Z-QG incorporated peptides. (A) Mass chromatograms for tryptic digests of Z-QG-treated and nontreated S-tag EGFP. Top column: peptide mass mapping of nontreated S-tag EGFP (base peak presentation). second-bottom columns: mass chromatograms for specific peptides, FER, ETAAAK and ETAAAK(Z-QG)FER, the m/z values of which are 451.2, 590.3, and 1342.5, respectively, showing the analyses of Z-QG-treated S-tag EGFP (solid lines) and nontreated S-tag EGFP (dotted lines). The relative ratio of peptide peak intensity for each analysis was adjusted using a common peptide, FEGDTLVNR [m/z 1051.1 (+1)]. (B and C) Mass spectra for ETAAAK(Z-QG)FER and GSGM*K(Z-QG)ETAAAK, respectively, with indication of singly and doubly charged ions. Met was identified in the oxidized form (M*).

Figure 3. Principle of the direct monitoring of MTG-catalyzed cross-linking of S-tag-EBFP and S-tag-EGFP through their N-terminal extra S-peptide sequence using FRET.

significant increase was observed without MTG, indicating the formation of a heterodimer between S-tag-EGFP and S-tag-EBFP, which permits us to detect the FRET signal in accordance with the cross-linking reaction. With respect to the kinetics of the TG-catalyzed reaction, competitive hydrolysis of Gln to Glu should be taken into account (28, 29) and it makes analysis of the reaction somewhat complicated. Since the FRET signal solely reflects the productive cross-linking reaction, our strategy offers a simple way to evaluate transglutamination without considering nonproductive hydrolysis. To verify the applicability of the FRET signal for following the cross-linking process of target proteins, we designed simple experiments by focusing our attention on the heterodimerization. If the molar ratio of S-tag-EBFP to the partner EGFP protein is changed from 1/1 to 4/1 or 1/4 while keeping the total protein concentration constant, the composition of the heterodimer will be theoretically altered from 50% to 32%. Consequently, the FRET signal is expected to level off at a lower saturation level than that of the equimolar mixture. As shown in Figure 5, the FRET signal obtained at the 4/1 or 1/4 ratios

Figure 4. Fluorescent spectral change in the presence (A) or the absence (B) of MTG with an equimolar mixture of S-tagEBFP and S-tag-EGFP.

was saturated at a lower level compared to that in the 1/1 ratio as we anticipated. Furthermore, the rate of FRET signal increase at the 4/1 or 1/4 ratios was about

Controllable Protein Cross-Linking by Transglutaminase

Figure 5. Time course of the cross-linking reaction between S-tag-EBFP and S-tag-EGFP followed by the change in FRET signal. The cross-linking reaction was carried out by varying the molar ratio of S-tag-EBFP to S-tag-EGFP at 1/1 (closed circles, with MTG; open triangles, without MTG), 1/4 (open circles, with MTG) and 4/1 (closed triangles, with MTG).

60-70% of that at the 1/1 ratio, being almost consistent with the theoretical value (64%). These results ensure that the signal increase reflected the cross-linking reaction between S-tag-EGFP and S-tag-EBFP catalyzed by MTG, indicating the feasibility of FRET-based evaluation of a protein cross-linking reaction. It seemed that the level of saturation of the FRET signal at the 4/1 ratio was slightly lower than that at the 1/4 ratio. This might be explained by the reabsorption of photons in the presence of an excess amount of S-tag-EBFP, which resulted in a decrease in the apparent number of S-tagEBFP that participated in the generation of the FRET signal. Controlling the Protein Cross-Linking by Engineering S-Peptide. The limited conjugation through an extra peptidyl linker of proteins of interest could lead us a stoichiometric preparation of dimeric proteins. As shown in Figure 1B (lane 5), a small portion of the tetrameric form of S-tag-EGFP was evident, suggesting that MTG catalyzes the subsequent cross-linking reaction of two dimers to a tetramer. As we identified two acyl acceptors, K5 and K11 (Figure 2), the formation of the tetramer could be ascribed to the presence of at least two types of dimeric forms of S-tag-EGFP, being distinctly cross-linked through Gln15-Lys5 or Gln15-Lys11. To gain further information on the tetramer formation, we prepared an engineered S-peptide of which Lys5, one of the previously identified acyl acceptors, was replaced by Ala, which results in the reduction of either the combination or the number of free Lys residues in dimeric forms of S-tag-EGFP. Then the cross-linking behavior of K5A-tagEGFP by MTG was compared with that of S-tag-EGFP (Figure 6). In both cases, the cross-linking reaction seemed to reach the equilibrium after 40 min based on SDS-PAGE analyses because no further change was observed in all protein bands. Figure 6A shows the accumulation of tetrameric forms of S-tag-EGFP concomitant with the formation of dimeric forms. On the other hand, in the case of K5A-tag-EGFP, the tetramer formation was significantly reduced by eliminating the possibility of Gln15-Lys5 cross-link (Figure 6B), suggesting that engineering S-peptide would facilitate the limited conjugation of target proteins approaching 1 to 1 stoichiometry. We attempted to estimate the relative portion of the

Bioconjugate Chem., Vol. 14, No. 2, 2003 355

Figure 6. Time course of the cross-linking reaction of S-tagEGFP (A) and K5A-tag-EGFP (B) followed by the SDS-PAGE analysis. Time zero indicates the intact samples without MTG in both cases.

tetramer to the dimer by the image analysis of Figure 6. The portion of the tetramer relative to the dimer was changed from approximately 3/7 with S-tag-EGFP to 1/9 with K5A-tag-EGFP, suggesting the selective dimer formation. However, slower cross-linking and more portion of the residual monomeric forms (due to the competitive hydrolysis of Gln to Glu (28) or the intramolecular cross-linking previously observed with guinea pig transglutaminase (27) as shown in Figure 1) were evident with K5A-tag-EGFP compared to S-tag-EGFP, implying the reduction of cross-linkability of the engineered S-peptide. More sophysticated protein cross-linking reaction requires further improvements of the peptidyl linker with regard to both the reactivity and the selectivity. It should be noted that only the dimeric and the tetrameric forms were observed with no detectable formation of the trimer in both cases, suggesting that a monomer reacted with a higher activity with another monomer than with a dimer. This might be explained if a monomer experiences greater steric constraint around the cross-linking site of a dimer. However, it is hard to imagine that the reactivity between dimers would exceed that between a monomer and a dimer by simply taking steric aspects into account. MTG might have an inherent property that prefers cross-linked polymeric products. Although we have not had clear answers for this phenomenon yet, on the basis of the results obtained here, one can at least conclude that the protein cross-linking by MTG can be controlled by manipulating a peptidyl linker. Controlling the Protein Cross-Linking by S-Protein. Protection and deprotection steps are common in multistep organic syntheses when target molecules have several reactive functional groups. For instance, a number of protective groups are employed in solid-phase peptide synthesis. When considering the TG-mediated conjugation of more than two proteins that possess several peptide linkers or susceptible sites for TG, there may be the possibility of temporal protection of a specific peptide tag for TG. S-protein, the partner protein of S-peptide for the formation of ribonuclease A (30), could be a perfect fit for this purpose. We thus examined the capability of S-protein for the suppression of the reactivity of the S-peptide for MTG. The effect of S-protein on the cross-linking reaction was evaluated by two means: the change in FRET signal and SDS-PAGE analysis.

356 Bioconjugate Chem., Vol. 14, No. 2, 2003

Figure 7. Suppression of the cross-linking reaction by capping S-peptide with S-protein confirmed by the FRET signal and SDS-PAGE analysis (closed circles, with MTG; open circles, with both MTG and S-protein). Inset: lane 1, low molecular weight markers; lane 2, S-tag-EGFP; lane 3, S-tag-EGFP treated with MTG; lane 4, S-tag-EGFP treated with S-protein; lane 5, S-tag-EGFP treated with S-protein and MTG.

Figure 7 clearly demonstrates that S-protein completely abolished the ability of S-peptide to act as a peptidyl substrate, suggesting for the first time controllable protein cross-linking of a target protein by TG. In this sense, S-protein is thus a potent candidate as a protective group in the protein conjugation by MTG through Speptide. The next step should be directed at the modulation of the specific interaction (i.e., deprotection). In fact, it has been reported that the affinity of S-peptide for S-protein could be tunable by site-directed mutagenesis (30). In our case, the effect of amino acid substitution on the reactivity toward MTG should be taken into account as well. In conclusion, our results demonstrate the potential of S-peptide as a novel peptidyl linker for tunable protein cross-linking by MTG. A new strategy that allows direct monitoring of the MTG-catalyzed protein cross-linking reaction using FRET enables us to optimize such a peptidyl linker. Our concept and experimental technique are promising for TG-directed enzymatic manipulation of recombinant proteins. Further studies are underway in our laboratory. ACKNOWLEDGMENT

We are grateful to Ajinomoto Co., Inc., for providing the MTG sample. We would also like to thank Drs. Naoto Tonouchi, Yoshiyuki Fujishima and Yoshiyuki Kumazawa (Ajinomoto Co. Inc.) for helpful discussion on the characteristics of MTG. The present work was supported by the Grant-in-Aid for Exploratory Research (13875156) from the Ministry of Education, Culture, Science, Sports and Technology of Japan. LITERATURE CITED (1) Hermanson, G. T. (1996) Functional targets. Bioconjugate Techniques, pp 3-23, Academic Press, San Diego. (2) Folk, E. J. (1983) Mechanism and basis for substrate specificity of transglutaminase-catalyzed -(γ-glutamyl)lysine bond formation. Adv. Enzymol. Relat. Areas Mol. Biol. 54, 1-56. (3) Wold, F. (1985) Reactions of the amide side-chains of glutamine and asparagine in vivo. Trends Biochem. Sci. 10, 4-6. (4) Coussons, P. J., Kelly, S. M., Price, N. C., Johnson, C. M., Smith, B., and Sawyer, L. (1991) Selective modification by

Kamiya et al. transglutaminase of a glutamine side chain in the hinge region of the histidine-388 f glutamine mutant of yeast phosphoglycerate kinase. Biochem. J. 273, 73-78. (5) Sato, H., Ikeda, M., Suzuki, K., and Hirayama, K. (1996) Site-specific modification of interleukin-2 by the combined use of genetic engineering technique and transglutaminase. Biochemistry 35, 13072-13080. (6) Ando H., Adachi M., Umeda K., Matsuura A., Nonaka M., Uchio R., Tanaka H., and Motoki M. (1989) Purification and characteristics of novel transglutaminase derived from microorganisms. Agric. Biol. Chem. 53, 2613-2617. (7) Kanaji, T., Ozaki, H., Takao, T., Kawajiri, H., Ide, H., Motoki, M., and Shimonishi, Y. (1993) Primary structure of microbial transglutaminase from Streptoverticillium sp. strain s-8112. J. Biol. Chem. 268, 11565-11572. (8) Sato, H., Hayashi, E., Yamada, N., Yatagai, M., Takahara, Y. (2001) Further studies on the site-specific protein modification by microbial transglutaminase. Bioconjugate Chem. 12, 701-710. (9) Motoki, M., and Seguro, K. (1998) Transglutaminase and its use for food processing. Trends Food Sci. Technol. 9, 204-210. (10) Kamata, Y., Ishikawa, E., and Motoki, M. (1992) Enzyme immobilization on ion exchangers by forming an enzyme coating with transglutaminase as a crosslinker. Biosci. Biotech. Biochem. 56, 1323-1324. (11) Josten, A., Meusel, M., and Spener, F. (1998) Microbial transglutaminase-mediated synthesis of hapten-protein conjugates for immunoassays. Anal. Biochem. 258, 202-208. (12) Josten, A., Haalk, L, Spener, F., and Meusel, M. (2000) Use of microbial transglutaminase for the enzymatic biotinylation of antibody. J. Immunol. Methods 240, 47-54. (13) Bechtold, U., Otterbach, J. T., Pasternack, R., and Fuchsbauer, H.-L. (2000) Enzymic preparation of protein G-peroxidase conjugates catalysed by transglutaminase. J. Biochem. 127, 239-245. (14) Ota, M., Sawa, A., Nio, N., and Ariyoshi, Y. (1999) Enzymatic ligation for synthesis of single-chain analogue of monellin by transglutaminase. Biopolymers 50, 193-200. (15) Wada, E., Matsuda, H., Imamura, K., Sakiyama, T., Adachi, S., Matsuno, R., Nakanishi, K. (2001) Enzymatic modification of β-lactoglobulin with N-fatty-acyl-dipeptide by transglutaminase from Streptomyces mobaraensis. Biotechnol. Lett. 23, 1367-1372. (16) Josten, A., Meusel, M., Spencer, F., and Haalk, L. (1999) Enzyme immobilization via microbial transglutaminase: a method for the generation of stable sencing surface. J. Mol. Catal. B. Enzymol. 7, 58-66. (17) Kamiya, N., Ogawa, T., Nagamune, T. (2001) Enhancement of apparent thermostability of lipase from Rhizopus sp. by the treatment with microbial transglutaminase. Biotechnol. Lett. 23, 1629-1632. (18) Nonaka, M., Matsuura, Y., and Motoki, M. (1996) Incorporation of lysine- and lysine-dipeptides into as1-casein by Ca2+-independent microbial transglutaminase. Biosci. Biotech. Biochem. 60, 131-133. (19) Matsumura, Y., Chanyongvorakul, Y., Kumazawa, Y., Ohtsuka, T., and Mori, T. (1996) Enhanced susceptibility to transglutaminase reaction of R-lactalbumin in the molten globule state. Biochim. Biophys. Acta 1292, 69-76. (20) Taguchi, S., Nishiyama, K., Igi, K., Ito, K., Taira, H., Motoki, M., and Momose, H. (2000) Substrate specificity analysis of microbial transglutaminase using proteinaceous protease inhibitors as natural model substrates. J. Biochem. 128, 415-425. (21) Seitz, A., Schneider, F., Pasternack, R., Fuchsbauer, H.L., and Hampp, N. (2001) Enzymatic cross-linking of purple membranes catalyzed by bacterial transglutaminase. Biomacromolecules 2, 233-238. (22) Folk, J. E. and Cole, P. W. (1966) Mechanism of action of guinea pig liver transglutaminase. I. Purification and properties of the enzyme: identification of a functional cystein essential for activity. J. Biol. Chem. 241, 5518-5525. (23) Arai, R., Ueda, H., Tsumoto, K., Mahoney, W. C., Kumagai, I., Nagamune, T. (2000) Fluorolabeling of antibody variable domains with green fluorescent protein variants: application

Controllable Protein Cross-Linking by Transglutaminase to an energy transfer-based homogeneous immunoassay. Protein Eng. 13, 369-376. (24) Tsien, R. Y. (1998) The green fluorescent protein. Annu. Rev. Biochem. 67, 509-544. (25) Suzuki, T., Terasaki, M., Takemoto-Hori, C., Hanada, T., Ueda, T., Wada, A. and Watanabe, K. (2001) Structural compensation for deficit of rRNA with proteins in mammalian mitochondrial ribosome; systematic analysis of protein components of the large ribosomal subunit from mammalian mitochondria. J. Biol. Chem. 276, 21724-21736. (26) Suzuki, T., Terasaki, M., Takemoto-Hori, C., Hanada, T., Ueda, T., Wada, A. and Watanabe, K. (2001) Proteomic analysis of the mammalian mitochondrial ribosome; Identification of protein components in the 28S small subunit. J. Biol. Chem. 276, 33181-33195.

Bioconjugate Chem., Vol. 14, No. 2, 2003 357 (27) Furutani, Y., Kato, A., Notoya, M., Ghoneim, M. A., Hirose, S. (2001) A simple assay and histochemical localization of transglutaminase activity using a derivative of green fluorescent protein as substrate. J. Histochem. Cytochem. 49, 247-258. (28) Gross, M., Whetzel, N. K., Folk, J. E. (1975) The extended active site of guinea pig liver transglutaminase. J. Biol. Chem. 250, 4648-4655. (29) Day, N., Keillor, J. W. (1999) A continuous spectrophotometric linked enzyme assay for transglutaminase activity. Anal. Biochem. 274, 141-144. (30) Kim, J.-S., Raines, R. T. (1993) Ribonuclease S-peptide as a carrier in fusion proteins. Protein Sci. 2, 348-356.

BC025610Y