- Email: [email protected]
Circular dichroism analysis of the glucan binding domain of Streptococcus mutans glucan binding protein-A
Biochimica et Biophysica Acta 1384 Ž1998. 112–120
Circular dichroism analysis of the glucan binding domain of Streptococcus mutans glucan binding protein-A Wolfgang Haas b
a,)
, Robert MacColl b, Jeffrey A. Banas
a
a Albany Medical College-A68, 47 New Scotland AÕe., Albany, NY 12208, USA Wadsworth Center, New York State Department of Health, P.O. Box 509, Albany, NY 12201, USA
Received 15 September 1997; revised 8 December 1997; accepted 15 December 1997
Abstract The glucan binding domain ŽGBD. of the glucan binding protein-A ŽGBP-A. from the cariogenic bacterium Streptococcus mutans was studied using circular dichroism ŽCD. analysis, Chou–Fasman–Rose secondary structure prediction, and absorption and fluorescence spectroscopy. Our data show that the binding domain undergoes a conformational shift upon binding to the ligand dextran. The CD spectrum shows two positive bands at 280 nm and 230 nm which were assigned to aromatic residues. The 230-nm band was seen at 208C and 308C, lost intensity at 408C, and was eliminated at 458C coinciding with complete denaturation. The protein was stable at physiological pH, but precipitated at pH 5. A pH of 10 changed the secondary structure but had no effect on the 230-nm band. Analysis of the CD data in the far UV using the SELCON computer program revealed a high content of b-sheets and a lack of a-helical structures. Secondary structure prediction based on the amino acid sequence of GBD agreed with the CD analysis. The fluorescence emission maximum at 339 nm suggested that the majority of the tryptophans were located in the interior of the protein. This maximum shifted to higher energy upon binding to the ligand dextran. q 1998 Elsevier Science B.V. Keywords: Glucan binding domain; Circular dichroism; Secondary structure; Ž Streptococcus mutans .
1. Introduction Commensal and pathogenic bacteria generally require some mechanism of adherence to the host in order to be able to colonize their ecological niche. This is especially true for bacteria that colonize the smooth surfaces of the teeth where the constant flow of saliva and the mechanical forces of the tongue and food movement make strong adhesion necessary w1x. Oral streptococci, such as Streptococcus mutans, se-
)
Corresponding author. Fax: q1-518-262-5748; E-mail: [email protected]
crete glucosyltransferases which use dietary sucrose to produce extracellular glucan polymers. These glucans are very adhesive by nature and allow the bacteria to form a biofilm on the tooth surface w1,2x. S. mutans was shown to possess at least three distinct, nonenzymatic glucan binding proteins ŽGBP. that are either secreted or cell-bound w3–5x. While the cell-surface protein GBP-C clearly has a lectin-like function w5x, the role of the other two proteins remains to be elucidated. GBP-A, formerly described as the 74-kDa GBP, was cloned and sequenced w3x and shown to bind to dextran-like glucans containing a-1,6-glucosidic linkages. Binding is mediated by a glucan binding domain ŽGBD. which is present at the
0167-4838r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 8 3 8 Ž 9 8 . 0 0 0 0 5 - 3
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
carboxyl-terminus of the protein and consists of a number of repeating units w3,6,7x. This domain has homologous counterparts in all glucosyltransferases from oral streptococci, the Toxins A and B from Clostridium difficile, and several muramidases from Streptococcus pneumoniae and its bacteriophages w6x. Determination of the structural basis underlying the protein–polysaccharide interactions would not only further our general understanding of the relationship between protein structure and function, but it could also lead to strategies to combat the infectious diseases caused by the organisms that express these proteins. Protein crystals for X-ray crystallography are not yet available whereas analysis by nuclear magnetic resonance ŽNMR. is not possible due to the high molecular weight of the proteins. The circular dichroism Ž CD. spectrum of a protein reveals information about the conformation of the peptide backbone and allows calculation of the contribution of a-helix, b-sheet, b-turn, and random coil to the secondary structure of the protein w8,9x. CD is very effective in monitoring changes in the secondary and tertiary structures in protein binding studies. Here we report the use of CD and absorption and fluorescence spectroscopy to study the structure of the GBP-A GBD under various conditions.
2. Materials and methods 2.1. Construction of fusion protein B:GBD The DNA region encoding the GBD beginning with amino acid 158 and proceeding to the carboxylterminus w3x of S. mutans GBP-A was amplified by polymerase chain reaction Ž PCR. , using the phosphorylated primers TCTCTCCAACCAATAGCTTCTTT and CCGCCATATTTACCGTTTTCAA. The PCR product was ligated into the EcoRV site of the PinPoint Xa1 vector Ž Promega. , yielding the plasmid pDPL4. Cloning and plasmid transformation were done using standard procedures w10x. The plasmid expressed, after induction with isopropyl-b-D-thiogalacto-pyranoside ŽIPTG., a fusion protein ŽB:GBD. between the biotinylated 1.3 S transcarboxylase subunit from Propionibacterium shermanii and the GBD. B:GBD reacted with an anti-GBP-A antibody on a Western immunoblot and had the same molecular
113
weight on an SDS-PAGE as deduced from the amino acid sequence Ž data not shown. . The advantages of the B:GBD fusion protein, compared to the native protein, were that it could be expressed in high amounts in Escherichia coli cells, it could be easily purified, and the GBD could be separated from the transcarboxylase subunit by specific protease treatment. Additionally, it was ascertained that the B:GBD fusion had virtually identical glucan binding properties as the native GBP-A. 2.2. Protein purification E. coli JM109 cells containing pDPL4 were grown at 378C in 2xYT w10x medium containing 100 m grml ampicillin and 2 m M biotin. Overnight cultures were diluted 1:100 in fresh medium and allowed to grow for 1 h before expression of the fusion protein was induced by adding IPTG to a final concentration of 100 m M. The cells were harvested after 4–5 h by centrifugation at 3000 = g for 20 min at 48C in a Sorvall GSA rotor. Cells were resuspended in 10 volumes Žwt.rvol.. TEN3 buffer Ž20 mM TRIS–HCl, pH 7.2, 1 mM EDTA, 300 mM NaCl, 100 m M phenylmethylsulfonyl fluoride. and disrupted by passing twice through a French pressure cell Ž 1 IN. DIA., model 4-3339, American Instrument, Silver Spring, MD. at a flow rate of 2 mlrmin at 20,000 P.S.I. Cellular debris was removed by centrifugation for 30 min at 10,000 = g at 48C in a Sorvall SS34 rotor. The supernatant was then incubated for 2 h at 48C with Sepharcyl S1000 ŽSigma. which acts as an affinity matrix for proteins containing a GBD. The matrix was washed extensively with TEN3 buffer and 0.5 M guanidinium hydrochloride. Bound protein was eluted with 6 M guanidinium hydrochloride and dialyzed at 48C overnight against 0.1 M phosphate buffer ŽpH 6.8. with multiple changes of buffer. This protein preparation was subsequently incubated for 2 h at 48C with SoftLink resin Ž Promega. . This resin consists of avidin, which specifically binds to biotinylated proteins such as B:GBD and therefore functions as a second affinity column. The matrix was washed with 0.1 M phosphate buffer and B:GBD was eluted with buffer containing 5 m M biotin. The protein was dialyzed overnight at 48C against 0.05 M phosphate buffer ŽpH 6.8 unless indicated otherwise. and used the following day for analysis.
114
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
2.3. CleaÕage of B:GBD with factor Xa The PinPoint vector was designed such that a recognition site for the protease factor Xa was placed at the carboxyl-terminus of the transcarboxylase subunit. This specific protease site allows separation of the two components of the fusion protein without further degrading the two resulting peptides. Digestion with factor Xa was performed as follows: 350 m g protein was digested with 14 m g protease at pH 7.4 in a total volume of 1.2 ml for 16 h at 258C. The transcarboxylase subunit was removed by passing over an avidin column and the remaining GBDpeptide was dialyzed against 0.05 M phosphate buffer. Around 6–8 m g of B:GBD or GBD were analyzed by SDS-PAGE gels, visualized using the Silver Stain Kit from Boehringer Mannheim Ž Indianapolis, IN. and found to be free of other proteins. 2.4. Circular dichroism analysis A JASCO J720 spectropolarimeter was used for CD analysis. The light path for the near UV region was 10 mm, and for the far UV region 0.5 mm was used. Measurements were taken in 0.05 M sodium phosphate buffer, pH 6.8. Temperature was kept at 208C Žunless indicated otherwise. with circulating water from a Neslab RTE III refrigerated circulator. Far-UV data were recorded within the wavelength range of 260 nm to 180 nm at 0.2-nm decrements with a scan speed of 20 nmrs, a 1.0-nm band width, and an averaging time of 0.5 s. A base line was taken with buffer or buffer containing dextran at the same conditions as used for the sample and subtracted from each protein spectrum. Six scans in the far UV region and four scans in the near UV-region were averaged on each sample to improve the signal-to-noise ratios. Noise was further reduced using a Savitzky–Golay filter. For data in the far UV region, molar units were calculated using a mean residue molecular weight of 111.4.
tivity of 1.93 for a 1 mgrml solution in a 1-cm light path. 2.6. Prediction of protein secondary structure The Self-Consistent Method of Protein Secondary Structure Estimation ŽSELCON. by Sreerama and Woody w11x was used to analyze the CD data. The methods of Kabsch and Sander Ž KS. w12x, Levitt and Greer ŽLG. w13x and Hennessey and Johnson ŽHJ. w14x were used for the assignment of secondary structure. 2.7. Absorption and fluorescence emission spectra The UV absorption spectrum was determined for B:GBD at 208C using a Beckman DU640 spectrophotometer. The fluorescence emission spectrum of B:GBD at 208C was obtained using a Perkin-Elmer fluorescence spectrophotometer LS50B. The excitation and emission slits were at 5 nm, and the excitation wavelength were 280 and 295 nm.
3. Results 3.1. CD analysis of the GBD in the presence and absence of dextran B:GBD was expressed and purified as described above. The GBD was released from the fusion protein by proteolysis and subjected to far-UV CD analysis. The corresponding CD spectrum revealed two positive bands at 230 and 202 nm, a negative band at 196 nm and a broad negative band between 210 and 215 nm ŽFig. 1. . Addition of dextran to yield a 10-fold molar excess over the concentration of GBD
2.5. Measurement of protein concentrations Protein concentrations were determined either with the BCA Protein Assay Ž Pierce, Rockford, IL. using bovine serum albumin as a standard or by measuring the absorbance at 277 nm and dividing by the absorp-
Fig. 1. Far-UV CD spectra of the GBD in the absence Žsolid. and presence of 30 m M dextran Ždotted..
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
resulted in an increase in ellipticity above 225 nm and a decreased ellipticity below 225 nm. The broad negative band shifted to 208–214 nm while the position of the peaks at 202 and 196 nm did not change, although their ellipticity decreased drastically. Since CD analysis in the far-UV spectrum yields information about the secondary structure of a protein, the observed change in the spectrum of GBD indicates that the GBD undergoes a conformational shift upon binding to dextran. The dextran-solution alone was not optically active and therefore not detected by CD. 3.2. CD analysis of the B:GBD fusion protein The B:GBD fusion protein, with the transcarboxylase subunit at the amino-terminus and the GBD at the carboxyl-terminus, was analyzed in the presence and absence of dextran. As seen with the GBD alone, the spectrum of B:GBD showed an intense positive band at 230 nm and a change due to the addition of dextran, indicating a change in protein conformation due to interactions with the polysaccharide Ž data not shown..
115
prediction of any a-helices or a decrease of b-sheets below 48% Ždata not shown.. Hennessey et al. w15x predicted 3% a-helices, 32% b-sheets and 18% bturns for the 1.3 S transcarboxylase subunit at pH 5.8 which changed only slightly at pH 9.0. Therefore, the presence of the transcarboxylase subunit did not result in drastic changes in secondary structure predictions from GBD to B:GBD. Our data indicate that GBD contains only b-sheets and no a-helices which has been shown to be the case for other lectins such as concanavalin A w16x. However, as will be discussed below, the high amount of aromatic residues in the two proteins and their contribution to the far-UV spectrum may or may not have obscured the secondary structure analysis w9,17x. Secondary structure predictions based on the amino acid sequence of the GBD using the methods of Chou and Fasman w18,19x and Rose w20x ŽMacDNASIS, Hitachi Software, San Bruno, CA. also came to the conclusion that GBD is rich in b-sheet structures ŽFig. 2., though 12–21% of the domain is predicted to have the potential to form either a-helices or b-sheets. 3.4. Analysis of B:GBD from 230 to 350 nm
3.3. Secondary structure predictions The computer program SELCON, using the assignments of Kabsch and Sander, was used to predict the secondary structures of GBD and B:GBD in the presence and absence of dextran. As seen in Table 1, neither protein seemed to contain a-helical structures, but consisted of at least 55% b-sheet. Using the assignments of Levitt and Greer, or Hennessey and Johnson, the ratio of b-sheets to b-turn and other structures changed but did not result in the Table 1 Secondary structure of GBD and B:GBD in the presence and absence of dextran as predicted by the SELCON computer program using the assignment of Kabsch and Sander GBD GBDqdextran B:GBD B:GBDqdextran
a-Helix a
b-Sheet
b-Turn
Other
y9.1 y0.5 y2.1 y1.4
61.8 55.3 58.2 55.1
21.1 23.0 22.3 24.7
23.3 22.0 17.6 22.0
Data are given in percent. a Negative values were introduced by the computer program for mathematical reasons and should be regarded as ‘zero’.
The CD spectrum of B:GBD was measured between 230 and 350 nm to study the contributions of the aromatic residues to the near CD spectrum. Disulfide bonds are known to contribute to the near-UV CD spectrum but were absent in the proteins studied w9,21x. As seen in Fig. 3, B:GBD gave rise to a positive band centered around 230 nm and a second positive band at 280 nm. Exciton splitting could not be shown to be responsible for this unusual spectrum since no negative band of equal intensity was observed in close proximity to either one of the positive bands w9,22x. Exciton splitting occurs when two or more chromophores are situated close together in the proper orientation, resulting in a CD spectrum that has negative and positive bands of equal rotational strength. Previous reports w9,17,23,24x showed that aromatic amino acids, especially tyrosine residues, can create a positive band centered around 220–250 nm that is not due to the folding of the peptide backbone. The presence in GBD of 38 Ž9.1%. tyrosine, 26 Ž 6.2%. phenylalanine, and 11 Ž2.6%. tryptophan residues out of a total of 417 amino acids, and the lack of cysteine residues, make it likely that these
116
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
Fig. 2. Secondary structure prediction for the GBD using the method of Chou, Fasman and Rose. Capital letters underneath an amino acid indicate that this residue is very likely to be present in an a-helix ŽH., b-sheet ŽS., b-turn ŽT. or random coil ŽC.. Lower case letters indicate that the amino acid has the potential to be part of this secondary structure.
residues are responsible for the positive band at 230 nm, as well as the band at 280 nm. Interestingly, 38 of these aromatic residues are arranged in 12 clusters, each containing three or more aromatic amino acids.
Fig. 3. CD spectrum of B:GBD at 230–350 nm.
3.5. Protein conformation at Õarious pH Day w23x showed for the G5 protein that a positive band at 228 nm was due to tyrosine residues based on the protein’s CD spectrum at various pH. We measured the CD spectrum of B:GBD between pH 5.0 and 11.5 Ž Fig. 4.. At pH 5.0, more than 75% of the protein precipitated, which is not surprising since the predicted p I for B:GBD is 5.0. The spectrum remained unchanged between pH 5.5 and 8.5, indicating that the GBD is stable at physiological pH. Increasing the pH to 10.0 resulted in an alteration of the spectrum below 225 nm while the 230 nm band remained unaffected. A pH of 11.5, which is well above the p K a of 10.13 for the hydroxyl group of
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
Fig. 4. Far UV CD spectra of B:GBD at pH 6 Žlight solid., pH 7 Žsolid., pH 8.5 Ždashed. and pH 10 Ždotted..
117
B:GBD loses some of its tertiary structure while maintaining most of its secondary structure. Binding studies using B:GBD preincubated at 40 and 508C showed that the former protein bound to dextran as well as the untreated B:GBD, while the latter did not bind at all. Furthermore, 40% of the protein heated above 458C precipitated, indicating that loss of native conformation results in loss of functionality and solubility Ždata not shown.. 3.7. Absorption and fluorescence spectra
tyrosine w25x, resulted in a total loss of the 230 nm band Ždata not shown. , which suggested that tyrosine was the cause of the 230 nm band seen in GBD and B:GBD. 3.6. Effects of temperature on B:GBD’s conformation To analyze the effects that elevated temperatures might have on protein folding, B:GBD was gradually heated to the indicated temperature and allowed to remain at that temperature for 10 min before measurements were taken ŽFig. 5. . The spectra remained unchanged at 208 and 308C, but the intensity of the 230-nm band decreased at 408C, while the spectrum below 220 nm changed only slightly. At 458C and above, the 230-nm band had disappeared and the spectrum below 220 nm changed significantly from that seen at lower temperatures. Increasing the temperature up to 808C did not result in drastic changes in the CD spectrum, indicating that most of the structural changes occurred at 458C Ždata not shown. . The protein did not regain its former conformation when the temperature was decreased from 408C or 458C back to lower temperatures. These data indicate that 408C marks a transition temperature where
Fig. 5. CD spectra of B:GBD at 208 and 308C Žsolid., 408C Ždotted. and 458C Ždashed.. The sample was cooled down to 308C and the CD spectrum was measured again Žlight solid..
The absorption maximum of B:GBD was at 277 nm ŽFig. 6A. . The absorptivity for a 1 mgrml solution was determined to be 1.93 in a 1-cm light path. The fluorescence emission maximum was at 339 nm ŽFig. 6B. . This maximum suggests that the majority of the tryptophan residues were in the interior of the protein w26x. The emission spectra were analogous for 280 and 295-nm excitations. The fluorescence emission maximum was slightly blue shifted in the presence of dextran Ž Fig. 6B.. Heating B:GBD to 458C for 15 min before fluorescence analysis did not change
Fig. 6. Absorption ŽA. and fluorescence ŽB. spectra of B:GBD. The fluorescence emission spectra of B:GBD were obtained using 295 nm excitation in the presence Žsolid. and absence Ždotted. of dextran.
118
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
the emission peak after excitation at 295 nm, indicating that the environment of the tryptophane-residues was unaltered.
4. Discussion The presence of a GBD in virulence factors of oral streptococci and homologous domains in several other proteins from pathogenic bacteria w6x led us to study the GBD from GBP-A. Rather than using the native protein, we chose to use a fusion protein between the 1.3 S transcarboxylase subunit of known CD spectrum and the GBD. This fusion protein was equal to the native protein in its ability to bind to dextran but had the advantage that the GBD could be separated from the transcarboxylase subunit by specific protease digestion. The opportunity to use two independent affinity columns to obtain highly purified protein preparations was a second advantage of this approach. The CD spectrum and proposed secondary structure distributions of the 1.3 S transcarboxylase subunit were published by Hennessey et al. w15x. The authors describe a negative band from 220–240 nm and negative ellipticity starting at 200 nm. The CD spectrum of B:GBD and GBD showed an unusual positive band at 230 nm that is believed to be due to the contribution of aromatic side chains. Day w23x showed that a positive band at 228 nm in the CD spectrum of the G5 protein from bacteriophage fd was due to phenolic transitions of tyrosine residues. Lytic amidase from S. pneumoniae and lysozyme from phage Cp-1 w27,28x contain C-terminal cholinebinding domains that are homologous to the GBD of GBP-A. The CD spectra of both proteins show a strong positive band around 230 nm that was attributed to the contribution of aromatic side chains. This was confirmed for B:GBD by measuring the CD spectrum at various pH, since Day w23x and Sanz and Garcia w28x could show that this band is pH dependent. Their results and our data demonstrate that the 230 nm band is lost at a pH of 11.5 or higher, which is sufficiently high to deprotonate the phenolic hydroxyl group. The protein is stable between pH 5.5 and 8.5, but precipitates at pH 5.0, which equals its predicted p I. The aromatic side chain contribution to the CD spectrum from 180 to 260 nm may or may not
obscure the secondary structure analysis as pointed out by Fasman w9x and Woody w17x. Multiple sets of CD data obtained for B:GBD and GBD in the presence or absence of dextran were analyzed using the SELCON computer program w11x and the assignments of Kabsch and Sander, Levitt and Greer, and Hennessey and Johnson. All assignments agreed that the GBD is made up of at least 48% b-sheets and contains no a-helices. Sanz and Garcia w28x used several methods to assess the secondary structure of the choline binding domain of CPL-1 lysozyme, a domain homologous to GBD. They concluded that the C-terminal domain is made up of two sheetbend-helix modules connected by five bend-sheet structures and that the binding to choline results in the redistribution of b-sheets and b-turns in lysozyme. However, definite structure assignments will have to await X-ray crystallography or NMR analysis of the GBD or of a homologous structure. Heat-denaturation experiments revealed that the GBD undergoes a transition at 408C. At this temperature, the peak at 230 nm loses in intensity while the remainder of the spectrum undergoes little or no change. This suggests that the changes affect only the tertiary structure that brings the aromatic residues in close contact, but not the protein’s secondary structure. If the temperature is raised to 458C, the CD spectrum loses the 230-nm band entirely, producing a change in the secondary structure of the protein as well. A secondary structure prediction using the method of Chou, Fasman and Rose revealed a large amount of b-sheets and that at least 12% of the GBD has the potential to form either a b-sheet or an a-helix. It is possible that all or some of these regions are in a b-sheet conformation in the native state and at elevated temperatures undergo a shift to an a-helical structure which is more stable, making the transition irreversible. Although we could not predict the secondary structure based on the CD spectrum of heat-denatured B:GBD, this theory is supported by data obtained for aged protein. The spectrum of 6-week old B:GBD had a spectrum similar to that obtained for heat-denatured protein Ždata not shown. in that it lacked a positive band at 230 nm and had a broad negative band between 205 and 225 nm. The SELCON computer program predicted 16% a-helix, 39% b-sheet, 24% b-turn and 23% other structures for this protein sample, which is in good agreement
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
with the Chou–Fasman–Rose data and would explain the conformational changes seen at elevated temperatures. The fluorescence emission maximum at 339 nm is characteristic of tryptophan residues found in the interior of proteins. When Trp residues are on the surface of proteins in the presence of water the Trp maxima are at lower energies. The emission spectra were similar using excitations at either 280 or 295 nm. This result suggests that only Trp residues are fluorescing since excitation at 280 nm excites both Trp and Tyr, whereas excitation at 295 nm excites Trp only. Fluorescence of Tyr would be observed at about 310 nm, and the absence of this band suggests that the Tyr emission is quenched by the Trp. The Trp emission was slightly blue-shifted in the presence of dextran. This suggests that some of the Trp residues were now even more shielded from water. This nonpolar environment could be caused by a conformational change w26x, which is in good agreement with the CD data. Much has been written about the GBD in GBP-A and the GTFs and the repeating units present in the GBD. However, these publications were restricted to defining homologies or correlating the size of the binding domain to the affinity for a ligand w6,7,29,30x. With the combination of Chou–Fasman–Rose secondary structure prediction and CD analysis we were able to assign a secondary structure for the GBD. Our data show that the repeating units form b-sheet structures that are interrupted by b-turn or random coil structures. The unusual positive band at 230 nm could be contributed to the high amount of Tyr residues in the protein. The contribution of certain highly conserved amino acids in forming the secondary structures and ligand binding is currently under investigation. Binding to dextran induces a conformational shift in the GBD that might allow a larger portion of the protein to interact with the ligand. More work is currently in progress in order to get a better understanding of how the GBD is structurally organized and binds to dextran.
Acknowledgements The authors thank L.E. Eisele ŽBiochemistry Core. for help with the SELCON computer analysis and
119
T.T. Andersen and L.A. Day for critical review of the manuscript. This research was funded by grant DE10058 from the National Institute of Dental Research.
References w1x S. Hamada, H.D. Slade, Biology, immunology, and cariogenicity of Streptococcus mutans, Biol. Rev. 44 Ž1980. 331–384. w2x W.L. Loesche, Role of Streptococcus mutans in human dental decay, Microbiol. Rev. 50 Ž1986. 353–380. w3x J.A. Banas, R.R.B. Russell, J.J. Ferretti, Sequence analysis of the gene for the glucan-binding protein of Streptococcus mutans Ingbritt, Infect. Immunol. 58 Ž1990. 667–673. w4x D.J. Smith, H. Akita, W.F. King, M.A. Taubman, Purification and antigenicity of a novel glucan-binding protein of Streptococcus mutans, Infect. Immunol. 62 Ž1994. 2545– 2552. w5x Y. Sato, Y. Yamamoto, H. Kizaki, Cloning and sequence analysis of the gbpC gene encoding a novel glucan-binding protein of Streptococcus mutans, Infect. Immunol. 65 Ž1997. 668–675. w6x P.M. Giffard, N.A. Jacques, Definition of a fundamental repeating unit in streptococcal glucosyltransferase glucan-binding regions and related sequences, J. Dent. Res. 73 Ž1994. 1133–1141. w7x C. von Eichel-Streiber, M. Sauerborn, H.K. Kuramitsu, Evidence for a modular structure of the homologous repetitive C-terminal carbohydrate-binding sites of Clostridium difficile toxins and Streptococcus mutans glucosyltransferases, J. Bacteriol. 174 Ž1992. 6707–6710. w8x N.J. Greenfield, Methods to estimate the conformation of proteins and polypeptides from circular dichroism data, Anal. Biochem. 235 Ž1996. 1–10. w9x G.D. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum, New York, 1996. w10x J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning, Cold Spring Harbor Laboratory Press, 1989. w11x N. Sreerama, R.W. Woody, A self-consistent method for the analysis of protein secondary structure from circular dichroism, Anal. Biochem. 209 Ž1993. 32–44. w12x W. Kabsch, C. Sander, Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features, Biopolymers 22 Ž1983. 2577–2637. w13x M. Levitt, J. Greer, Automatic identification of secondary structure in globular proteins, J. Mol. Biol. 114 Ž1977. 181–203. w14x J.P. Hennessey Jr., W.C. Johnson, Information content in the circular dichroism of proteins, Biochemistry 20 Ž1981. 1085–1094. w15x J.P. Hennessey Jr., W.C. Johnson, C. Bahler, H.G. Wood, Subunit interactions of transcarboxylase as studied by circular dichroism, Biochemistry 21 Ž1982. 642–646.
120
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
w16x F.A. Quiocho, Carbohydrate-binding proteins: tertiary structures and protein–sugar interactions, Annu. Rev. Biochem. 55 Ž1986. 287–315. w17x R.W. Woody, Aromatic side-chain contributions to the far ultraviolet circular dichroism of peptides and proteins, Biopolymers 17 Ž1978. 1451–1467. w18x P.Y. Chou, G.D. Fasman, Prediction of the secondary structure of proteins from their amino acid sequence, Adv. Enzymol. 47 Ž1978. 45–148. w19x P.Y. Chou, G.D. Fasman, Empirical predictions of protein conformation, Annu. Rev. Biochem. 47 Ž1978. 251–276. w20x G.D. Rose, Prediction of chain turns in globular proteins on a hydrophobic basis, Nature 272 Ž1978. 586–590. w21x J.L. Garcia, E. Diaz, A. Romero, P. Garcia, Carboxyl-terminal deletion analysis of the major pneumococcal autolysin, J. Bacteriol. 176 Ž1994. 4066–4072. w22x N. Ito, T. Takagi, Circular dichroism of crystalline L-cysteine in potassium bromide disc. Possibility of exciton splitting, Biochim. Biophys. Acta 221 Ž1970. 430–441. w23x L.A. Day, Circular dichroism and ultraviolet absorption of a deoxyribonucleic acid binding protein of filamentous bacteriophage, Biochemistry 12 Ž1973. 5329–5339. w24x M.C. Manning, R.W. Woody, Theoretical study of the
w25x
w26x w27x
w28x
w29x
w30x
contribution of aromatic side chains to the circular dichroism of basic bovine pancreatic trypsin inhibitor, Biochemistry 28 Ž1989. 8609–8613. R.M.C. Dawson, D.C. Elliott, W.H. Elliott, K.M. Jones ŽEds.., Data for Biochemical Research, Oxford Univ. Press, New York, 1969. D. Freifelder, Physical Biochemistry, 2nd edn., Freeman, New York, 1982. J.M. Sanchez-Puelles, J.M. Sanz, J.L. Garcia, E. Garcia, Cloning and expression of gene fragments encoding the choline-binding domain of pneumococcal murein hydrolases, Gene 89 Ž1990. 69–75. J.M. Sanz, J.L. Garcia, Structural studies of the lysozyme coded by the pneumococcal phage Cp-1, Eur. J. Biochem. 187 Ž1990. 409–416. C. Kato, H.K. Kuramitsu, Carboxyl-terminal deletion analysis of the Streptococcus mutans glucosyltransferase-I enzyme, FEMS Microbiol. Lett. 72 Ž1990. 299–302. M. Lis, T. Shiroza, H.K. Kuramitsu, Role of the C-terminal direct repeating units of the Streptococcus mutans glucosyltransferase-S in glucan binding, Appl. Environ. Microbiol. 61 Ž1995. 2040–2042.
Circular dichroism analysis of the glucan binding domain of Streptococcus mutans glucan binding protein-A Wolfgang Haas b
a,)
, Robert MacColl b, Jeffrey A. Banas
a
a Albany Medical College-A68, 47 New Scotland AÕe., Albany, NY 12208, USA Wadsworth Center, New York State Department of Health, P.O. Box 509, Albany, NY 12201, USA
Received 15 September 1997; revised 8 December 1997; accepted 15 December 1997
Abstract The glucan binding domain ŽGBD. of the glucan binding protein-A ŽGBP-A. from the cariogenic bacterium Streptococcus mutans was studied using circular dichroism ŽCD. analysis, Chou–Fasman–Rose secondary structure prediction, and absorption and fluorescence spectroscopy. Our data show that the binding domain undergoes a conformational shift upon binding to the ligand dextran. The CD spectrum shows two positive bands at 280 nm and 230 nm which were assigned to aromatic residues. The 230-nm band was seen at 208C and 308C, lost intensity at 408C, and was eliminated at 458C coinciding with complete denaturation. The protein was stable at physiological pH, but precipitated at pH 5. A pH of 10 changed the secondary structure but had no effect on the 230-nm band. Analysis of the CD data in the far UV using the SELCON computer program revealed a high content of b-sheets and a lack of a-helical structures. Secondary structure prediction based on the amino acid sequence of GBD agreed with the CD analysis. The fluorescence emission maximum at 339 nm suggested that the majority of the tryptophans were located in the interior of the protein. This maximum shifted to higher energy upon binding to the ligand dextran. q 1998 Elsevier Science B.V. Keywords: Glucan binding domain; Circular dichroism; Secondary structure; Ž Streptococcus mutans .
1. Introduction Commensal and pathogenic bacteria generally require some mechanism of adherence to the host in order to be able to colonize their ecological niche. This is especially true for bacteria that colonize the smooth surfaces of the teeth where the constant flow of saliva and the mechanical forces of the tongue and food movement make strong adhesion necessary w1x. Oral streptococci, such as Streptococcus mutans, se-
)
Corresponding author. Fax: q1-518-262-5748; E-mail: [email protected]
crete glucosyltransferases which use dietary sucrose to produce extracellular glucan polymers. These glucans are very adhesive by nature and allow the bacteria to form a biofilm on the tooth surface w1,2x. S. mutans was shown to possess at least three distinct, nonenzymatic glucan binding proteins ŽGBP. that are either secreted or cell-bound w3–5x. While the cell-surface protein GBP-C clearly has a lectin-like function w5x, the role of the other two proteins remains to be elucidated. GBP-A, formerly described as the 74-kDa GBP, was cloned and sequenced w3x and shown to bind to dextran-like glucans containing a-1,6-glucosidic linkages. Binding is mediated by a glucan binding domain ŽGBD. which is present at the
0167-4838r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 8 3 8 Ž 9 8 . 0 0 0 0 5 - 3
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
carboxyl-terminus of the protein and consists of a number of repeating units w3,6,7x. This domain has homologous counterparts in all glucosyltransferases from oral streptococci, the Toxins A and B from Clostridium difficile, and several muramidases from Streptococcus pneumoniae and its bacteriophages w6x. Determination of the structural basis underlying the protein–polysaccharide interactions would not only further our general understanding of the relationship between protein structure and function, but it could also lead to strategies to combat the infectious diseases caused by the organisms that express these proteins. Protein crystals for X-ray crystallography are not yet available whereas analysis by nuclear magnetic resonance ŽNMR. is not possible due to the high molecular weight of the proteins. The circular dichroism Ž CD. spectrum of a protein reveals information about the conformation of the peptide backbone and allows calculation of the contribution of a-helix, b-sheet, b-turn, and random coil to the secondary structure of the protein w8,9x. CD is very effective in monitoring changes in the secondary and tertiary structures in protein binding studies. Here we report the use of CD and absorption and fluorescence spectroscopy to study the structure of the GBP-A GBD under various conditions.
2. Materials and methods 2.1. Construction of fusion protein B:GBD The DNA region encoding the GBD beginning with amino acid 158 and proceeding to the carboxylterminus w3x of S. mutans GBP-A was amplified by polymerase chain reaction Ž PCR. , using the phosphorylated primers TCTCTCCAACCAATAGCTTCTTT and CCGCCATATTTACCGTTTTCAA. The PCR product was ligated into the EcoRV site of the PinPoint Xa1 vector Ž Promega. , yielding the plasmid pDPL4. Cloning and plasmid transformation were done using standard procedures w10x. The plasmid expressed, after induction with isopropyl-b-D-thiogalacto-pyranoside ŽIPTG., a fusion protein ŽB:GBD. between the biotinylated 1.3 S transcarboxylase subunit from Propionibacterium shermanii and the GBD. B:GBD reacted with an anti-GBP-A antibody on a Western immunoblot and had the same molecular
113
weight on an SDS-PAGE as deduced from the amino acid sequence Ž data not shown. . The advantages of the B:GBD fusion protein, compared to the native protein, were that it could be expressed in high amounts in Escherichia coli cells, it could be easily purified, and the GBD could be separated from the transcarboxylase subunit by specific protease treatment. Additionally, it was ascertained that the B:GBD fusion had virtually identical glucan binding properties as the native GBP-A. 2.2. Protein purification E. coli JM109 cells containing pDPL4 were grown at 378C in 2xYT w10x medium containing 100 m grml ampicillin and 2 m M biotin. Overnight cultures were diluted 1:100 in fresh medium and allowed to grow for 1 h before expression of the fusion protein was induced by adding IPTG to a final concentration of 100 m M. The cells were harvested after 4–5 h by centrifugation at 3000 = g for 20 min at 48C in a Sorvall GSA rotor. Cells were resuspended in 10 volumes Žwt.rvol.. TEN3 buffer Ž20 mM TRIS–HCl, pH 7.2, 1 mM EDTA, 300 mM NaCl, 100 m M phenylmethylsulfonyl fluoride. and disrupted by passing twice through a French pressure cell Ž 1 IN. DIA., model 4-3339, American Instrument, Silver Spring, MD. at a flow rate of 2 mlrmin at 20,000 P.S.I. Cellular debris was removed by centrifugation for 30 min at 10,000 = g at 48C in a Sorvall SS34 rotor. The supernatant was then incubated for 2 h at 48C with Sepharcyl S1000 ŽSigma. which acts as an affinity matrix for proteins containing a GBD. The matrix was washed extensively with TEN3 buffer and 0.5 M guanidinium hydrochloride. Bound protein was eluted with 6 M guanidinium hydrochloride and dialyzed at 48C overnight against 0.1 M phosphate buffer ŽpH 6.8. with multiple changes of buffer. This protein preparation was subsequently incubated for 2 h at 48C with SoftLink resin Ž Promega. . This resin consists of avidin, which specifically binds to biotinylated proteins such as B:GBD and therefore functions as a second affinity column. The matrix was washed with 0.1 M phosphate buffer and B:GBD was eluted with buffer containing 5 m M biotin. The protein was dialyzed overnight at 48C against 0.05 M phosphate buffer ŽpH 6.8 unless indicated otherwise. and used the following day for analysis.
114
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
2.3. CleaÕage of B:GBD with factor Xa The PinPoint vector was designed such that a recognition site for the protease factor Xa was placed at the carboxyl-terminus of the transcarboxylase subunit. This specific protease site allows separation of the two components of the fusion protein without further degrading the two resulting peptides. Digestion with factor Xa was performed as follows: 350 m g protein was digested with 14 m g protease at pH 7.4 in a total volume of 1.2 ml for 16 h at 258C. The transcarboxylase subunit was removed by passing over an avidin column and the remaining GBDpeptide was dialyzed against 0.05 M phosphate buffer. Around 6–8 m g of B:GBD or GBD were analyzed by SDS-PAGE gels, visualized using the Silver Stain Kit from Boehringer Mannheim Ž Indianapolis, IN. and found to be free of other proteins. 2.4. Circular dichroism analysis A JASCO J720 spectropolarimeter was used for CD analysis. The light path for the near UV region was 10 mm, and for the far UV region 0.5 mm was used. Measurements were taken in 0.05 M sodium phosphate buffer, pH 6.8. Temperature was kept at 208C Žunless indicated otherwise. with circulating water from a Neslab RTE III refrigerated circulator. Far-UV data were recorded within the wavelength range of 260 nm to 180 nm at 0.2-nm decrements with a scan speed of 20 nmrs, a 1.0-nm band width, and an averaging time of 0.5 s. A base line was taken with buffer or buffer containing dextran at the same conditions as used for the sample and subtracted from each protein spectrum. Six scans in the far UV region and four scans in the near UV-region were averaged on each sample to improve the signal-to-noise ratios. Noise was further reduced using a Savitzky–Golay filter. For data in the far UV region, molar units were calculated using a mean residue molecular weight of 111.4.
tivity of 1.93 for a 1 mgrml solution in a 1-cm light path. 2.6. Prediction of protein secondary structure The Self-Consistent Method of Protein Secondary Structure Estimation ŽSELCON. by Sreerama and Woody w11x was used to analyze the CD data. The methods of Kabsch and Sander Ž KS. w12x, Levitt and Greer ŽLG. w13x and Hennessey and Johnson ŽHJ. w14x were used for the assignment of secondary structure. 2.7. Absorption and fluorescence emission spectra The UV absorption spectrum was determined for B:GBD at 208C using a Beckman DU640 spectrophotometer. The fluorescence emission spectrum of B:GBD at 208C was obtained using a Perkin-Elmer fluorescence spectrophotometer LS50B. The excitation and emission slits were at 5 nm, and the excitation wavelength were 280 and 295 nm.
3. Results 3.1. CD analysis of the GBD in the presence and absence of dextran B:GBD was expressed and purified as described above. The GBD was released from the fusion protein by proteolysis and subjected to far-UV CD analysis. The corresponding CD spectrum revealed two positive bands at 230 and 202 nm, a negative band at 196 nm and a broad negative band between 210 and 215 nm ŽFig. 1. . Addition of dextran to yield a 10-fold molar excess over the concentration of GBD
2.5. Measurement of protein concentrations Protein concentrations were determined either with the BCA Protein Assay Ž Pierce, Rockford, IL. using bovine serum albumin as a standard or by measuring the absorbance at 277 nm and dividing by the absorp-
Fig. 1. Far-UV CD spectra of the GBD in the absence Žsolid. and presence of 30 m M dextran Ždotted..
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
resulted in an increase in ellipticity above 225 nm and a decreased ellipticity below 225 nm. The broad negative band shifted to 208–214 nm while the position of the peaks at 202 and 196 nm did not change, although their ellipticity decreased drastically. Since CD analysis in the far-UV spectrum yields information about the secondary structure of a protein, the observed change in the spectrum of GBD indicates that the GBD undergoes a conformational shift upon binding to dextran. The dextran-solution alone was not optically active and therefore not detected by CD. 3.2. CD analysis of the B:GBD fusion protein The B:GBD fusion protein, with the transcarboxylase subunit at the amino-terminus and the GBD at the carboxyl-terminus, was analyzed in the presence and absence of dextran. As seen with the GBD alone, the spectrum of B:GBD showed an intense positive band at 230 nm and a change due to the addition of dextran, indicating a change in protein conformation due to interactions with the polysaccharide Ž data not shown..
115
prediction of any a-helices or a decrease of b-sheets below 48% Ždata not shown.. Hennessey et al. w15x predicted 3% a-helices, 32% b-sheets and 18% bturns for the 1.3 S transcarboxylase subunit at pH 5.8 which changed only slightly at pH 9.0. Therefore, the presence of the transcarboxylase subunit did not result in drastic changes in secondary structure predictions from GBD to B:GBD. Our data indicate that GBD contains only b-sheets and no a-helices which has been shown to be the case for other lectins such as concanavalin A w16x. However, as will be discussed below, the high amount of aromatic residues in the two proteins and their contribution to the far-UV spectrum may or may not have obscured the secondary structure analysis w9,17x. Secondary structure predictions based on the amino acid sequence of the GBD using the methods of Chou and Fasman w18,19x and Rose w20x ŽMacDNASIS, Hitachi Software, San Bruno, CA. also came to the conclusion that GBD is rich in b-sheet structures ŽFig. 2., though 12–21% of the domain is predicted to have the potential to form either a-helices or b-sheets. 3.4. Analysis of B:GBD from 230 to 350 nm
3.3. Secondary structure predictions The computer program SELCON, using the assignments of Kabsch and Sander, was used to predict the secondary structures of GBD and B:GBD in the presence and absence of dextran. As seen in Table 1, neither protein seemed to contain a-helical structures, but consisted of at least 55% b-sheet. Using the assignments of Levitt and Greer, or Hennessey and Johnson, the ratio of b-sheets to b-turn and other structures changed but did not result in the Table 1 Secondary structure of GBD and B:GBD in the presence and absence of dextran as predicted by the SELCON computer program using the assignment of Kabsch and Sander GBD GBDqdextran B:GBD B:GBDqdextran
a-Helix a
b-Sheet
b-Turn
Other
y9.1 y0.5 y2.1 y1.4
61.8 55.3 58.2 55.1
21.1 23.0 22.3 24.7
23.3 22.0 17.6 22.0
Data are given in percent. a Negative values were introduced by the computer program for mathematical reasons and should be regarded as ‘zero’.
The CD spectrum of B:GBD was measured between 230 and 350 nm to study the contributions of the aromatic residues to the near CD spectrum. Disulfide bonds are known to contribute to the near-UV CD spectrum but were absent in the proteins studied w9,21x. As seen in Fig. 3, B:GBD gave rise to a positive band centered around 230 nm and a second positive band at 280 nm. Exciton splitting could not be shown to be responsible for this unusual spectrum since no negative band of equal intensity was observed in close proximity to either one of the positive bands w9,22x. Exciton splitting occurs when two or more chromophores are situated close together in the proper orientation, resulting in a CD spectrum that has negative and positive bands of equal rotational strength. Previous reports w9,17,23,24x showed that aromatic amino acids, especially tyrosine residues, can create a positive band centered around 220–250 nm that is not due to the folding of the peptide backbone. The presence in GBD of 38 Ž9.1%. tyrosine, 26 Ž 6.2%. phenylalanine, and 11 Ž2.6%. tryptophan residues out of a total of 417 amino acids, and the lack of cysteine residues, make it likely that these
116
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
Fig. 2. Secondary structure prediction for the GBD using the method of Chou, Fasman and Rose. Capital letters underneath an amino acid indicate that this residue is very likely to be present in an a-helix ŽH., b-sheet ŽS., b-turn ŽT. or random coil ŽC.. Lower case letters indicate that the amino acid has the potential to be part of this secondary structure.
residues are responsible for the positive band at 230 nm, as well as the band at 280 nm. Interestingly, 38 of these aromatic residues are arranged in 12 clusters, each containing three or more aromatic amino acids.
Fig. 3. CD spectrum of B:GBD at 230–350 nm.
3.5. Protein conformation at Õarious pH Day w23x showed for the G5 protein that a positive band at 228 nm was due to tyrosine residues based on the protein’s CD spectrum at various pH. We measured the CD spectrum of B:GBD between pH 5.0 and 11.5 Ž Fig. 4.. At pH 5.0, more than 75% of the protein precipitated, which is not surprising since the predicted p I for B:GBD is 5.0. The spectrum remained unchanged between pH 5.5 and 8.5, indicating that the GBD is stable at physiological pH. Increasing the pH to 10.0 resulted in an alteration of the spectrum below 225 nm while the 230 nm band remained unaffected. A pH of 11.5, which is well above the p K a of 10.13 for the hydroxyl group of
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
Fig. 4. Far UV CD spectra of B:GBD at pH 6 Žlight solid., pH 7 Žsolid., pH 8.5 Ždashed. and pH 10 Ždotted..
117
B:GBD loses some of its tertiary structure while maintaining most of its secondary structure. Binding studies using B:GBD preincubated at 40 and 508C showed that the former protein bound to dextran as well as the untreated B:GBD, while the latter did not bind at all. Furthermore, 40% of the protein heated above 458C precipitated, indicating that loss of native conformation results in loss of functionality and solubility Ždata not shown.. 3.7. Absorption and fluorescence spectra
tyrosine w25x, resulted in a total loss of the 230 nm band Ždata not shown. , which suggested that tyrosine was the cause of the 230 nm band seen in GBD and B:GBD. 3.6. Effects of temperature on B:GBD’s conformation To analyze the effects that elevated temperatures might have on protein folding, B:GBD was gradually heated to the indicated temperature and allowed to remain at that temperature for 10 min before measurements were taken ŽFig. 5. . The spectra remained unchanged at 208 and 308C, but the intensity of the 230-nm band decreased at 408C, while the spectrum below 220 nm changed only slightly. At 458C and above, the 230-nm band had disappeared and the spectrum below 220 nm changed significantly from that seen at lower temperatures. Increasing the temperature up to 808C did not result in drastic changes in the CD spectrum, indicating that most of the structural changes occurred at 458C Ždata not shown. . The protein did not regain its former conformation when the temperature was decreased from 408C or 458C back to lower temperatures. These data indicate that 408C marks a transition temperature where
Fig. 5. CD spectra of B:GBD at 208 and 308C Žsolid., 408C Ždotted. and 458C Ždashed.. The sample was cooled down to 308C and the CD spectrum was measured again Žlight solid..
The absorption maximum of B:GBD was at 277 nm ŽFig. 6A. . The absorptivity for a 1 mgrml solution was determined to be 1.93 in a 1-cm light path. The fluorescence emission maximum was at 339 nm ŽFig. 6B. . This maximum suggests that the majority of the tryptophan residues were in the interior of the protein w26x. The emission spectra were analogous for 280 and 295-nm excitations. The fluorescence emission maximum was slightly blue shifted in the presence of dextran Ž Fig. 6B.. Heating B:GBD to 458C for 15 min before fluorescence analysis did not change
Fig. 6. Absorption ŽA. and fluorescence ŽB. spectra of B:GBD. The fluorescence emission spectra of B:GBD were obtained using 295 nm excitation in the presence Žsolid. and absence Ždotted. of dextran.
118
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
the emission peak after excitation at 295 nm, indicating that the environment of the tryptophane-residues was unaltered.
4. Discussion The presence of a GBD in virulence factors of oral streptococci and homologous domains in several other proteins from pathogenic bacteria w6x led us to study the GBD from GBP-A. Rather than using the native protein, we chose to use a fusion protein between the 1.3 S transcarboxylase subunit of known CD spectrum and the GBD. This fusion protein was equal to the native protein in its ability to bind to dextran but had the advantage that the GBD could be separated from the transcarboxylase subunit by specific protease digestion. The opportunity to use two independent affinity columns to obtain highly purified protein preparations was a second advantage of this approach. The CD spectrum and proposed secondary structure distributions of the 1.3 S transcarboxylase subunit were published by Hennessey et al. w15x. The authors describe a negative band from 220–240 nm and negative ellipticity starting at 200 nm. The CD spectrum of B:GBD and GBD showed an unusual positive band at 230 nm that is believed to be due to the contribution of aromatic side chains. Day w23x showed that a positive band at 228 nm in the CD spectrum of the G5 protein from bacteriophage fd was due to phenolic transitions of tyrosine residues. Lytic amidase from S. pneumoniae and lysozyme from phage Cp-1 w27,28x contain C-terminal cholinebinding domains that are homologous to the GBD of GBP-A. The CD spectra of both proteins show a strong positive band around 230 nm that was attributed to the contribution of aromatic side chains. This was confirmed for B:GBD by measuring the CD spectrum at various pH, since Day w23x and Sanz and Garcia w28x could show that this band is pH dependent. Their results and our data demonstrate that the 230 nm band is lost at a pH of 11.5 or higher, which is sufficiently high to deprotonate the phenolic hydroxyl group. The protein is stable between pH 5.5 and 8.5, but precipitates at pH 5.0, which equals its predicted p I. The aromatic side chain contribution to the CD spectrum from 180 to 260 nm may or may not
obscure the secondary structure analysis as pointed out by Fasman w9x and Woody w17x. Multiple sets of CD data obtained for B:GBD and GBD in the presence or absence of dextran were analyzed using the SELCON computer program w11x and the assignments of Kabsch and Sander, Levitt and Greer, and Hennessey and Johnson. All assignments agreed that the GBD is made up of at least 48% b-sheets and contains no a-helices. Sanz and Garcia w28x used several methods to assess the secondary structure of the choline binding domain of CPL-1 lysozyme, a domain homologous to GBD. They concluded that the C-terminal domain is made up of two sheetbend-helix modules connected by five bend-sheet structures and that the binding to choline results in the redistribution of b-sheets and b-turns in lysozyme. However, definite structure assignments will have to await X-ray crystallography or NMR analysis of the GBD or of a homologous structure. Heat-denaturation experiments revealed that the GBD undergoes a transition at 408C. At this temperature, the peak at 230 nm loses in intensity while the remainder of the spectrum undergoes little or no change. This suggests that the changes affect only the tertiary structure that brings the aromatic residues in close contact, but not the protein’s secondary structure. If the temperature is raised to 458C, the CD spectrum loses the 230-nm band entirely, producing a change in the secondary structure of the protein as well. A secondary structure prediction using the method of Chou, Fasman and Rose revealed a large amount of b-sheets and that at least 12% of the GBD has the potential to form either a b-sheet or an a-helix. It is possible that all or some of these regions are in a b-sheet conformation in the native state and at elevated temperatures undergo a shift to an a-helical structure which is more stable, making the transition irreversible. Although we could not predict the secondary structure based on the CD spectrum of heat-denatured B:GBD, this theory is supported by data obtained for aged protein. The spectrum of 6-week old B:GBD had a spectrum similar to that obtained for heat-denatured protein Ždata not shown. in that it lacked a positive band at 230 nm and had a broad negative band between 205 and 225 nm. The SELCON computer program predicted 16% a-helix, 39% b-sheet, 24% b-turn and 23% other structures for this protein sample, which is in good agreement
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
with the Chou–Fasman–Rose data and would explain the conformational changes seen at elevated temperatures. The fluorescence emission maximum at 339 nm is characteristic of tryptophan residues found in the interior of proteins. When Trp residues are on the surface of proteins in the presence of water the Trp maxima are at lower energies. The emission spectra were similar using excitations at either 280 or 295 nm. This result suggests that only Trp residues are fluorescing since excitation at 280 nm excites both Trp and Tyr, whereas excitation at 295 nm excites Trp only. Fluorescence of Tyr would be observed at about 310 nm, and the absence of this band suggests that the Tyr emission is quenched by the Trp. The Trp emission was slightly blue-shifted in the presence of dextran. This suggests that some of the Trp residues were now even more shielded from water. This nonpolar environment could be caused by a conformational change w26x, which is in good agreement with the CD data. Much has been written about the GBD in GBP-A and the GTFs and the repeating units present in the GBD. However, these publications were restricted to defining homologies or correlating the size of the binding domain to the affinity for a ligand w6,7,29,30x. With the combination of Chou–Fasman–Rose secondary structure prediction and CD analysis we were able to assign a secondary structure for the GBD. Our data show that the repeating units form b-sheet structures that are interrupted by b-turn or random coil structures. The unusual positive band at 230 nm could be contributed to the high amount of Tyr residues in the protein. The contribution of certain highly conserved amino acids in forming the secondary structures and ligand binding is currently under investigation. Binding to dextran induces a conformational shift in the GBD that might allow a larger portion of the protein to interact with the ligand. More work is currently in progress in order to get a better understanding of how the GBD is structurally organized and binds to dextran.
Acknowledgements The authors thank L.E. Eisele ŽBiochemistry Core. for help with the SELCON computer analysis and
119
T.T. Andersen and L.A. Day for critical review of the manuscript. This research was funded by grant DE10058 from the National Institute of Dental Research.
References w1x S. Hamada, H.D. Slade, Biology, immunology, and cariogenicity of Streptococcus mutans, Biol. Rev. 44 Ž1980. 331–384. w2x W.L. Loesche, Role of Streptococcus mutans in human dental decay, Microbiol. Rev. 50 Ž1986. 353–380. w3x J.A. Banas, R.R.B. Russell, J.J. Ferretti, Sequence analysis of the gene for the glucan-binding protein of Streptococcus mutans Ingbritt, Infect. Immunol. 58 Ž1990. 667–673. w4x D.J. Smith, H. Akita, W.F. King, M.A. Taubman, Purification and antigenicity of a novel glucan-binding protein of Streptococcus mutans, Infect. Immunol. 62 Ž1994. 2545– 2552. w5x Y. Sato, Y. Yamamoto, H. Kizaki, Cloning and sequence analysis of the gbpC gene encoding a novel glucan-binding protein of Streptococcus mutans, Infect. Immunol. 65 Ž1997. 668–675. w6x P.M. Giffard, N.A. Jacques, Definition of a fundamental repeating unit in streptococcal glucosyltransferase glucan-binding regions and related sequences, J. Dent. Res. 73 Ž1994. 1133–1141. w7x C. von Eichel-Streiber, M. Sauerborn, H.K. Kuramitsu, Evidence for a modular structure of the homologous repetitive C-terminal carbohydrate-binding sites of Clostridium difficile toxins and Streptococcus mutans glucosyltransferases, J. Bacteriol. 174 Ž1992. 6707–6710. w8x N.J. Greenfield, Methods to estimate the conformation of proteins and polypeptides from circular dichroism data, Anal. Biochem. 235 Ž1996. 1–10. w9x G.D. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum, New York, 1996. w10x J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning, Cold Spring Harbor Laboratory Press, 1989. w11x N. Sreerama, R.W. Woody, A self-consistent method for the analysis of protein secondary structure from circular dichroism, Anal. Biochem. 209 Ž1993. 32–44. w12x W. Kabsch, C. Sander, Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features, Biopolymers 22 Ž1983. 2577–2637. w13x M. Levitt, J. Greer, Automatic identification of secondary structure in globular proteins, J. Mol. Biol. 114 Ž1977. 181–203. w14x J.P. Hennessey Jr., W.C. Johnson, Information content in the circular dichroism of proteins, Biochemistry 20 Ž1981. 1085–1094. w15x J.P. Hennessey Jr., W.C. Johnson, C. Bahler, H.G. Wood, Subunit interactions of transcarboxylase as studied by circular dichroism, Biochemistry 21 Ž1982. 642–646.
120
W. Haas et al.r Biochimica et Biophysica Acta 1384 (1998) 112–120
w16x F.A. Quiocho, Carbohydrate-binding proteins: tertiary structures and protein–sugar interactions, Annu. Rev. Biochem. 55 Ž1986. 287–315. w17x R.W. Woody, Aromatic side-chain contributions to the far ultraviolet circular dichroism of peptides and proteins, Biopolymers 17 Ž1978. 1451–1467. w18x P.Y. Chou, G.D. Fasman, Prediction of the secondary structure of proteins from their amino acid sequence, Adv. Enzymol. 47 Ž1978. 45–148. w19x P.Y. Chou, G.D. Fasman, Empirical predictions of protein conformation, Annu. Rev. Biochem. 47 Ž1978. 251–276. w20x G.D. Rose, Prediction of chain turns in globular proteins on a hydrophobic basis, Nature 272 Ž1978. 586–590. w21x J.L. Garcia, E. Diaz, A. Romero, P. Garcia, Carboxyl-terminal deletion analysis of the major pneumococcal autolysin, J. Bacteriol. 176 Ž1994. 4066–4072. w22x N. Ito, T. Takagi, Circular dichroism of crystalline L-cysteine in potassium bromide disc. Possibility of exciton splitting, Biochim. Biophys. Acta 221 Ž1970. 430–441. w23x L.A. Day, Circular dichroism and ultraviolet absorption of a deoxyribonucleic acid binding protein of filamentous bacteriophage, Biochemistry 12 Ž1973. 5329–5339. w24x M.C. Manning, R.W. Woody, Theoretical study of the
w25x
w26x w27x
w28x
w29x
w30x
contribution of aromatic side chains to the circular dichroism of basic bovine pancreatic trypsin inhibitor, Biochemistry 28 Ž1989. 8609–8613. R.M.C. Dawson, D.C. Elliott, W.H. Elliott, K.M. Jones ŽEds.., Data for Biochemical Research, Oxford Univ. Press, New York, 1969. D. Freifelder, Physical Biochemistry, 2nd edn., Freeman, New York, 1982. J.M. Sanchez-Puelles, J.M. Sanz, J.L. Garcia, E. Garcia, Cloning and expression of gene fragments encoding the choline-binding domain of pneumococcal murein hydrolases, Gene 89 Ž1990. 69–75. J.M. Sanz, J.L. Garcia, Structural studies of the lysozyme coded by the pneumococcal phage Cp-1, Eur. J. Biochem. 187 Ž1990. 409–416. C. Kato, H.K. Kuramitsu, Carboxyl-terminal deletion analysis of the Streptococcus mutans glucosyltransferase-I enzyme, FEMS Microbiol. Lett. 72 Ž1990. 299–302. M. Lis, T. Shiroza, H.K. Kuramitsu, Role of the C-terminal direct repeating units of the Streptococcus mutans glucosyltransferase-S in glucan binding, Appl. Environ. Microbiol. 61 Ž1995. 2040–2042.