Báo cáo khoa học: EspB from enterohaemorrhagic Escherichia coli is a natively partially folded protein pptx

EspB from enterohaemorrhagic Escherichia coli is a natively partially folded protein Daizo Hamada 1 , Tomoaki Kato 1,2 , Takahisa Ikegami 3 , Kayo N. Suzuki 1 , Makoto Hayashi 2 , Yoshikatsu Murooka 2 , Takeshi Honda 4 and Itaru Yanagihara 1 1 Department of Developmental Infectious Diseases, Research Institute, Osaka Medical Center for Maternal and Child Health, Japan 2 Department of Biotechnology, Graduate School of Engineering, Osaka University, Japan 3 Laboratory of Structural Proteomics, Institute for Protein Research, Osaka University, Japan 4 Department of Bacterial Infections, Research Institute for Microbial Diseases, Osaka University, Japan Several bacteria, including enterohaemorrhagic and enteropathogenic Escherichia coli (EHEC and EPEC, respectively), express type III secretion systems [1] consisting of various proteins encoded at the genetic locus of enterocyte effacement [2–5]. To date, type III secretion systems have been identified in more than 20 pathogenic bacterial species [6]. Type III secretion systems are multiprotein complexes that span the bacterial and host membranes, permitting the direct delivery of effector proteins, such as the EPEC pro- teins [7], Tir [8–10], EspF [11,12], EspG [13] and Orf19 [14]. In the case of EHEC and EPEC, such complexes are formed by proteins including EspA, EspB and EspD [15,16]. Thus, the type III system regulates effector secretion and delivery into host cells. Keywords circular dichroism; natively partially folded proteins; nuclear magnetic resonance; fluorescence quenching; multiangle laser light scattering Correspondence I. Yanagihara, Department of Developmental Infectious Diseases, Research Institute, Osaka Medical Center for Maternal and Child Health, 840 Murodo, Izumi, Osaka 594-1011, Japan Fax: +81 725 57 3021 Tel: +81 725 56 1220 (ext. 5302) E-mail: itaruy@mch.pref.osaka.jp (Received 20 August 2004, revised 17 November 2004, accepted 2 December 2004) doi:10.1111/j.1742-4658.2004.04513.x The structural properties of EspB, a virulence factor of the Escherichia coli O157 type III secretion system, were characterized. Far-UV and near-UV CD spectra, recorded between pH 1.0 and pH 7.0, show that the protein assumes a-helical structures and that some tyrosine tertiary contacts may exist. All tyrosine side-chains are exposed to water, as determined by acryl- amide fluorescence quenching spectroscopy. An increase in the fluorescence intensity of 8-anilinonaphthalene-1-sulfonate was observed at pH 2.0 in the presence of EspB, whereas no such increase in fluorescence was observed at pH 7.0. These data suggest the formation of a molten globule state at pH 2.0. Destabilization of EspB at low pH was shown by urea-unfolding transitions, monitored by far-UV CD spectroscopy. The result from a sedi- mentation equilibrium study indicated that EspB assumes a monomeric form at pH 7.0, although its Stokes radius (estimated by multiangle laser light scattering) was twice as large as expected for a monomeric globular structure of EspB. These data suggest that EspB, at pH 7.0, assumes a relatively expanded conformation. The chemical shift patterns of EspB 15 N- 1 H heteronuclear single quantum correlation spectra at pH 2.0 and 7.0 are qualitatively similar to that of urea-unfolded EspB. Taken together, the properties of EspB reported here provide evidence that EspB is a natively partially folded protein, but with less exposed hydrophobic surface than traditional molten globules. This structural feature of EspB may be advan- tageous when EspB interacts with various biomolecules during the bacterial infection of host cells. Abbreviations ANS, 8-anilinonaphthalene-1-sulfonate; EHEC, enterohaemorrhagic Escherichia coli; EPEC, enteropathogenic Escherichia coli; HSQC, heteronuclear single quantum correlation; LB, Luria–Bertani. 756 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS EspB is an E. coli type III system protein that inter- acts with various biomolecules. For example, EspB binds to EspD, forming a pore complex of 3–5 nm diameter in the host cell membrane [17]. The N-ter- minal region of EspB also binds to host cell a-catenin and inhibits F-actin accumulation at adherence sites [18]. It has been recently shown that a1-antitrypsin, a host cellular protein, binds to and interferes with the function of EspB [19]. Moreover, EspB may bind to the external end of the filamentous apparatus formed by EspA proteins [20]. The filamentous apparatus is characteristic of type III secretion systems [21,22]. Fila- mentous EspA may form a conduit for translocation of bacterial effector proteins into host cells [23]. It has been suggested that EspA filaments attach to host cells via EspB ⁄ D pore complexes and that the pore complex also interacts specifically with the host protein, a-cate- nin [16]. However, other studies have demonstrated that EspB is not required for the interaction of EspA with host cells [20]. Although the precise functions of EspB during bac- terial infection are still somewhat ambiguous, the information discussed above indicates that EspB is a multifunctional protein with the potential to interact with various biological molecules. Knowledge of the conformational properties of EspB may clarify the role of EspB in bacterial attachment, but no information about the structural properties of EspB is currently available. In this study, we characterized the conformational properties of EspB in solution by using several spectro- scopic and hydrodynamic techniques, including CD, 8-anilinonaphthalene-1-sulfonate (ANS) binding, ultra- centrifugation, multiangle laser light scattering and heteronuclear NMR. The results of our analyses allow us to understand the conformational property of EspB and predict its role in bacterial infection to the host cell. Results CD The secondary structure of EspB, predicted from its amino acid sequence by using the PredictProtein server [24–26], indicates that the protein is predominantly a-helical (Fig. 1). As stated in the Experimental proce- dures, the recombinant EspB was purified from both soluble and insoluble fractions of cell lysates. At pH 7.0 and at a temperature of 20 °C, recombinant EspB prepared from the insoluble fraction showed a far-UV CD spectrum equivalent to EspB prepared from the soluble fraction. This suggests that both purification procedures adequately yielded the native conformation of EspB. The CD spectra are typical for the presence of a-helices (Fig. 2). However, the a-heli- cal content estimated from far-UV CD data is % 23%, Fig. 1. Secondary structure prediction of EspB derived from its amino acid sequence. H and E refer to a-helical and b-strand struc- tures, respectively. The data were obtained by using the Predict- Protein server [24,25]. Fig. 2. CD spectra of EspB. (A) Far-UV and (B) near-UV CD spectra of recombinant EspB purified from the insoluble fraction at pH 2.0 (dashed lines) and 7.0 (solid lines), and from the soluble fraction at pH 7.0 (s). (C) The dependence of the ellipticity, at 222 nm, on pH. D. Hamada et al. EspB is a natively partially folded protein FEBS Journal 272 (2005) 756–768 ª 2005 FEBS 757 which is substantially less than the predicted amount (76.3%; Table 1). The near-UV CD spectrum of EspB at pH 7.0 and 20 °C shows a minimum at around 280 nm. It is in the near-UV spectrum that aromatic residues display opti- cal activity. EspB contains three tyrosines at positions 66, 75 and 212, and no tryptophans. Therefore, the shape of the near-UV CD spectrum of EspB implies the presence of some tertiary contacts involving at least one of the tyrosines, although the intensity of each peak is not very high. To gain further insight into the conformational prop- erties of EspB, we recorded the far-UV CD spectrum of recombinant EspB prepared by different protocols between pH 1.0 and pH 7.0 (Fig. 2C). Interestingly, irrespective of the preparation procedures and pH con- ditions, these far-UV CD spectra are almost identical. Therefore, the amount of secondary structure seems to be virtually same at each pH value (Fig. 2 and Table 1). On the other hand, the near-UV CD spec- trum at pH 2.0 showed a less intense signal at 280 nm relative to the spectrum at pH 7.0, suggesting destabili- zation of tertiary interactions upon decreasing pH. Owing to the small difference observed here between the recombinant proteins prepared by the different procedures, we mostly used the EspB prepared from insoluble fraction because the purification yield was much higher. Quenching of protein tyrosine fluorescence by acrylamide A fluorescence spectrum of intrinsic tryptophan and tyrosine residues in proteins can be a good conforma- tional probe. In particular, the fluorescence quenching effect by small chemicals such as acrylamide provides information on the solvent-exposure of aromatic side- chains in proteins. As mentioned above, EspB contains only three tyro- sines and no tryptophan. The quenching effect of acryl- amide on the fluorescence of these EspB tyrosine side-chains at pH 7.0 was analyzed. Interestingly, a plot of F 0 ⁄ F obs vs. [Q] (Stern–Volmer plot [27]), where F 0 and F obs are the fluorescence intensities in the absence and presence of quencher, respectively, and [Q] is the concentration of quencher, shows a positive devi- ation from linearity at high acrylamide concentrations (Fig. 3). Therefore, the quenching behavior does not follow the simple Stern–Volmer equation (F 0 ⁄ F obs ¼ 1+K sv [Q]). This finding suggests that the tyrosine residues in EspB behave as independent fluorophores, each having their own K sv value. Additional informa- tion was obtained by analyzing the data using the following modified Stern–Volmer equation [28]: Table 1. Secondary structure content of EspB at various pH values. Values were calculated using the data from Fig. 2A in conjunction with the CDPRO package [70,71]. Predicted values are calculated from the results of secondary structure prediction (Fig. 1) using the PHDsec algorithm available at the PredictProtein server (http:// cubic.bioc.columbia.edu/predictprotein/) [24,25]. Condition a-Helix (%) b-Sheet (%) Others (%) pH 1.0 26.2 ± 1.8 20.5 ± 1.1 53.4 ± 0.9 pH 2.0 22.8 ± 1.5 22.9 ± 1.7 53.5 ± 0.8 pH 3.0 27.5 ± 1.9 20.0 ± 1.7 52.5 ± 1.0 pH 4.0 31.1 ± 0.5 16.9 ± 0.9 52.0 ± 1.0 pH 5.0 26.6 ± 2.1 20.8 ± 1.9 52.5 ± 0.4 pH 6.0 23.5 ± 1.4 22.8 ± 1.8 53.3 ± 0.8 pH 7.0 23.1 ± 1.1 22.9 ± 1.2 53.1 ± 0.9 Predicted values 76.3 4.2 19.5 Fig. 3. Fluorescence quenching of intrinsic tyrosine. (A) Stern–Volmer plot. (B) Modified Stern–Volmer plot (Eqn 1). EspB is a natively partially folded protein D. Hamada et al. 758 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS F 0 =ðF 0 À F obs Þ¼1=ðf a K sv ½QÞ þ 1=f a ð1Þ where f a is the fraction of accessible tyrosines. The plot of F 0 ⁄ (F 0 – F obs ) vs. 1 ⁄ [Q] (Fig. 3) shows a linear cor- relation between F 0 ⁄ (F 0 – F obs ) and 1 ⁄ [Q]. The values for K sv and f a are calculated as 32.7 ± 1.5Æm )1 and 1.05 ± 0.01, respectively. An f a value close to 1 sug- gests that the three tyrosine residues are likely to be solvent-exposed at neutral pH. ANS binding ANS binds to solvent-accessible hydrophobic surfaces and, when bound, an increase in ANS fluorescence intensity near 500 nm occurs. This property of ANS is often used to detect the presence of partially folded protein intermediates, e.g. molten globules [29]. Molten globule is originally defined as the partially folded state of protein that assumes a significant amount of native- like secondary structures but disrupted in tertiary struc- tures [30–38]. We used ANS fluorescence spectroscopy to probe the hydrophobic surface accessibility of EspB. At pH 4 and 7, the ANS fluorescence is low (Fig. 4), suggesting that hydrophobic surfaces are not exposed to solvent. On the other hand, ANS fluores- cence increases as the pH is decreased to 2.0 (Fig. 4). This observation suggests that hydrophobic surfaces are solvent-exposed at more acidic pH values. Under the same conditions, the protein assumes an a-helical conformation according to the far-UV CD spectrum at pH 2.0 (Fig. 2). The results obtained by CD and ANS fluorescence suggest the formation of a typical molten globule structure for EspB at acidic pH 2.0. Below pH 2.0, the ANS fluorescence decreased. For these experiments, the pH of the solution was adjusted by the addition of HCl. The decreased fluorescence intensity may be caused by the quenching effect of chloride ions on ANS fluorescence, rather than reflect- ing additional conformational changes in EspB. Urea unfolding Urea-induced unfolding transitions of EspB were monitored by far-UV CD spectroscopy. Plots of [h] at 222 nm vs. urea concentration show co-operative unfolding transitions throughout the pH range of 1.0–7.3 (Fig. 5). Between pH 3.0 and 7.3, unfolding Fig. 4. 8-Anilinonaphthalene-1-sulfonate (ANS) fluorescence at 500 nm as a function of pH. Data were taken at 20 °C in the pres- ence of 0.1 mgÆmL )1 EspB. Circles represent the raw data. The line is drawn only for visual assistance and is not a mathematical fit. Fig. 5. Urea unfolding of EspB at various pH values and at 20 °C. (A) The far-UV CD spectra obtained in the presence and absence of urea. The numbers refer to the concentration of added urea. (B) The urea-unfolding transition curves obtained at pH 2.0 (s), pH 5.4 (h) and pH 7.3 (n). Continuous lines are theoretical curves. The dotted and dashed lines correspond to the baselines for unfolded and folded states. D. Hamada et al. EspB is a natively partially folded protein FEBS Journal 272 (2005) 756–768 ª 2005 FEBS 759 transitions occur between 2.5 and 4.5 m urea, but shift to lower urea concentrations of 1.0–3.5 m at pH 2.0. The urea-induced unfolding curves (Fig. 5) were analyzed assuming a linear relationship between DG and urea concentration and assuming a two-state fold- ing mechanism, although there is no direct evidence that the transitions are two-state in nature. The derived parameters, DG water and m, are summarized in Table 2. Compared to the conformational state at pH 2.0, those at higher pH values are stabilized by % 3–10 kJÆ mol )1 . However, their m-values, which probably cor- relate with changes in the solvent-exposed surface area associated with unfolding (DASA), are similar regardless of pH. In our experiments, DASA largely reflects structural changes in the folded species. There- fore, given the m-values, EspB, at pH 2.0, which pos- sesses molten globule-like properties, has a similar accessible surface area as EspB conformations existing at higher pH. Therefore, EspB at around neutral pH should be a less compact structure than typical glob- ular proteins. Hydrodynamic property of EspB The hydrodymanic property of EspB has been ana- lyzed by multiangle dynamic scattering and ultracen- trifugation. Multiangle laser light scattering experiments for EspB at pH 2.0 and pH 7.0 revealed the presence of a single species with a Stokes radius of 3.7 and 3.1 nm, respectively (Fig. 6). A similar value was obtained at pH 4.0 and pH 6.0 (3.4 and 3.5 nm, respectively). This size is larger than the expected value for a globular protein of 32 kDa molecular mass, and corresponds to the value of globular pro- teins, of % 70 kDa. From its amino acid sequence, the molecular mass of the recombinant EspB is calcu- Table 2. Values of DG water and m for urea-induced unfolding of EspB. pH (kJÆmol )1 ) DG water (kJÆmol )1 ÆM )1 ) m 2.0 6.5 ± 2.1 3.9 ± 1.1 3.0 13.2 ± 1.7 3.9 ± 0.5 4.1 14.7 ± 1.9 4.1 ± 0.5 5.4 16.5 ± 1.6 3.8 ± 3.4 6.6 11.5 ± 2.3 3.4 ± 0.6 7.3 9.5 ± 1.3 3.0 ± 0.4 Fig. 6. Hydrodynamic property of EspB at 20 °C. Multiangle laser light scattering of EspB at pH 2.0 (A) and pH 7.0 (B). (C) Sedi- mentation equilibrium of EspB at pH 7.0. The data were analyzed assuming a single species in solution. In the lower panel, raw data are shown by circles, and the line is the theoretical curve. The upper panel shows the difference between raw data and theoretical values. EspB is a natively partially folded protein D. Hamada et al. 760 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS lated to be 32 kDa. Thus, if EspB assumes a rigid globular conformation at these conditions, the protein should assume a dimeric structure. The result of the sedimentation equilibrium study at pH 7.0 also indicated the presence of only a single spe- cies (Fig. 6C). The molecular mass estimated from this experiment is, however, 34 kDa, which is similar to the expected value for the monomeric EspB. From these results, we concluded that EspB at pH 7.0 assumes a relatively expanded monomeric con- formation whose Stokes radius is approximately twice as large as expected for a globular protein with a molecular weight similar to that of EspB. Importantly, the Stokes radius of EspB estimated from light scattering was almost independent of pro- tein concentration or pH value. This suggests that only a single monomeric species is present in each protein solution at pH 2.0–7.0. Heteronuclear NMR spectroscopy To further probe the structural properties of EspB, we recorded its 15 N- 1 H heteronuclear single quantum cor- relation (HSQC) spectra at pH 2.0 in the absence of urea and at pH 7.0 in the presence and absence of urea. At pH 2.0, in the absence of urea, the 15 N- 1 H HSQC spectrum shows little chemical shift dispersion (Fig. 7). Although the resolution is poor owing to the overlapping of peaks, the number of peaks that cor- respond to the main-chain 1 H- 15 N crosspeaks was estimated to be % 120. These peaks are relatively sharp and may reflect the amino acid residues that rapidly fluctuate with a timescale of nanosecond order. The recombinant EspB used in this study contains 333 amino acid residues. Thus, % 64% of main-chain 1 H- 15 N crosspeaks are missing. A similar phenomenon is often found for the molten globule state, reflecting the slow fluctuation of a particular region of protein molecules with a timescale of micro- second to millisecond order. This is highly consistent with our other spectroscopic studies, which show that the protein, at pH 2.0, is in partially folded confor- mation, similar to that of molten globules. Interest- ingly, the NMR spectrum of EspB at pH 7.0 in the absence of urea also shows little chemical shift disper- sion, with % 110 possible main-chain 1 H- 15 N cros- speaks. Thus, about 67% of main-chain 1 H- 15 N crosspeaks are probably slowly fluctuating. Both of the aforementioned NMR spectra are similar to the spectrum obtained for urea-unfolded EspB at pH 7.0. In this case, the number of peaks that correspond to main-chain 1 H- 15 N crosspeaks slightly increased to about 140. This implies the significant overlapping of crosspeaks or the presence of some residual struc- tures, even in the presence of 8.0 m urea. The result of little chemical shift dispersions with the small number of observable crosspeaks in the 15 N- 1 H HSQC spectrum of EspB at pH 7.0 in the absence of urea is inconsistent with the previous data obtained by CD and fluorescence spectroscopies showing the pres- ence of well-ordered conformations. This discrepancy suggests that, at neutral pH, EspB assumes a natively partially folded conformation without exposed hydro- phobic clusters accessible to ANS molecules. Fig. 7. 15 N- 1 H Heteronuclear single quantum correlation (HSQC) spectra of EspB taken at 15 °C. (A) pH 7.0 in the absence of urea. (B) pH 2.0 in the absence of urea. (C) pH 7.0 in the presence of 8.0 M urea. D. Hamada et al. EspB is a natively partially folded protein FEBS Journal 272 (2005) 756–768 ª 2005 FEBS 761 Discussion Conformational properties of EspB We analyzed the conformational properties of EspB by using three spectroscopic techniques. The spectral results suggest that EspB assumes intrinsically partially folded conformations [39–44] under various conditions. The different conformational states of EspB, found under different pH conditions, are summarized in Table 3. The shapes of the far-UV CD spectra suggest the presence of a substantial amount of secondary structure for EspB throughout the pH range of 1.0– 7.0. On the other hand, ANS fluorescence spectroscopy indicates that EspB shows a conformational transition involving the exposure of hydrophobic clusters when the pH of the protein solution is decreased to 2.0. Thus, at pH 2.0, the structure of EspB is consistent with the traditional definition of a molten globule, i.e. a compact partially folded state with a significant amount of native-like secondary structures, but disrup- ted in tertiary contacts [30–38]. Importantly, all EspB NMR spectra had chemical shift signals that were less dispersed than those of globular proteins, even at pH 7.0 in the absence of de- naturant. The NMR spectra reported here are qualita- tively similar to those observed for proteins that are unstructured when in the presence of denaturants. On the other hand, when analyzed by far-UV CD spectro- scopy, EspB showed the presence of secondary structures. EspB is therefore in a partially folded conformation, even at near-physiological conditions. However, ANS fluorescence spectroscopy suggests the presence of a negligible amount of exposed hydropho- bic surface for EspB at pH 7.0. This is a very unusual result because partially folded proteins generally have exposed hydrophobic clusters that are detected by increases of ANS fluorescence intensity. One possible explanation for the discrepancy could be that the hydrophobic clusters, found for the EspB molten glob- ule at pH 2.0, are disrupted in the structure found at pH 7.0. A similar situation is, indeed, often found for a-helical polypeptides in alcohol ⁄ water solvents [45–51]. However, this explanation can be ruled out as EspB is more stable at higher pH, which would prob- ably be inconsistent with the loss of intramolecular hydrophobic contacts. Thus, most EspB hydrophobic clusters should be buried at neutral pH. Variations in the conformational and thermodynamic properties of molten globules have been characterized. For example, the thermal unfolding experiments on the molten glob- ule state of a-lactalbumin shows a gradual transition, which suggests less organized hydrophobic contacts [34]. However, the cytochrome c molten globule state is highly ordered and the thermal unfolding transition of this species is co-operative with a clear enthalpy change upon unfolding [52,53]. This indicates that some organized hydrophobic contacts exist in the mol- ten globule state of cytochrome c. Furthermore, the presence of tertiary contacts in the molten globule states are shown by apomyoglobin and cytochrome c [54,55], and EspB also showed the presence of weak, but distinctive, peaks in the near-UV CD spectrum. Therefore, EspB, at neutral pH, may have the charac- ter of a highly ordered molten globule [56] with dis- tinct and ordered regions probably stabilized by the interactions between hydrophobic clusters. On the other hand, the NMR data also indicate the presence of highly fluctuating regions in EspB. As the data from ultracentrifugation and laser light scattering suggest that EspB assumes an expanded monomeric form, EspB may assume a partially folded structure with well-ordered regions and highly fluctuating regions under near-physiological conditions. Uversky et al. [41] proposed that natively unfolded proteins tend to have a low mean hydrophobicity and a relatively high net charge, and provided the following expression of inequality, <H><(<R> +1.151) ⁄ 2.785, between the hydrophobicity value <H> and the mean net charge <R> for this class of proteins. According to the amino acid sequence of EspB, its <H> and <R> values are 0.478 and 0.013, respectively. These values actually do not satisfy the above criteria. Several www servers, which predict the disordered regions in a protein from its amino acid sequence, are currently available. We used GlobPlot (http://globplot. embl.de/cgiDict.py), DisEMBL (http://dis.embl.de/ cgiDict.py) and PONDRÒ (http://www.pondr.com) [57–60]. In the case of DisEMBL, some disordered regions are predicted and, according to Remark-465 definition, residues at 12–27, 124–145, 157–188, 246– 257 and 303–312 are disordered. On the other hand, GlobPlot, when using the Russel ⁄ Linding definition, does not show a high probability for EspB to be largely Table 3. The conformational properties of EspB at pH 2.0 and 7.0, at a temperature of 20 °C. ANS, 8-anilinonaphthalene-1-sulfonate; HSQC, heteronuclear single quantum correlation. pH Far-UV CD Near-UV CD Hydrophobic exposure by ANS 15 N- 1 H HSQC Urea unfolding 7.0 Folded Folded Less exposed Unfolded Co-operative 2.0 Folded Partially folded Highly exposed Unfolded Co-operative EspB is a natively partially folded protein D. Hamada et al. 762 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS disordered. The results from PONDRÒ suggest that the amino acid residues at 1–53, 128–230 and 247–287 may be disordered. This is relatively consistent with the prediction of DisEMBL. If the prediction from PONDR is correct, only the amino acid sequences at residues 54–127 and 288–325 of EspB, i.e. one-third of the EspB sequence, assume ordered conformations. This amount may overestimate the disordered regions as the CD spectum indicates that about 50% of the EspB sequence should be, at least, partially folded. Of various approaches, only PONDR and DisEM- BL indicated that EspB may be natively partially folded. Incidentially, both prediction methods are based on artificial neural networks, whereas GlobPlot or the category shown by Uversky et al. [41] rely on the physicochemical propensities of amino acids to favor the disordered or globular structures. These results may not be surprising as, in contrast to natively unfolded proteins in general, EspB has a relatively ordered conformation. Importantly, human a-lactalbu- min, in the absence of Ca 2+ [34], adopts a typical mol- ten globule structure at neutral pH. However, none of the algorithims predict such a property of human a-lactalbumin. Thus, the prediction of natively parti- ally folded protein from its amino acid sequence should still be difficult compared with the prediction of natively unfolded proteins. Implications for the function of EspB It is well established that proteins fold to their unique native conformations, as determined by their amino acid sequences [61]. However, it is also clear that some proteins are unable to maintain well-defined structures, even under physiological conditions [39–44]. These proteins are often called natively unfolded or intrinsic- ally disordered proteins and assume either partially folded or completely unfolded conformations in an aqueous environment at neutral pH and, ideally, under near-physiological conditions. Our results clearly indicate that the structural char- acteristics of EspB are those of a natively partially folded protein. The far-UV CD spectra of IpaC, a homolog of EspB from Shigella flexneri, revealed an absence of significant amounts of secondary structure at neutral pH [62]. Thus, the intrinsically less organ- ized conformations of EspB and IpaC may be a com- mon property for this class of proteins. Importantly, some proteins that are natively unfol- ded show dramatic conformational changes into well- ordered structures when bound to their target molecules [40–44]. Therefore, it will be important to characterize the conformational state of EspB when bound to its target molecules, e.g. EspA, EspD, a-cate- nin and a1-antitrypsin. Using the genomic sequence of E. coli, Dunker and co-workers predicted that 8% of all proteins will have intrinsically disordered segments of greater than 50 res- idues in length [62]. Interestingly, the same predictions indicated that this percentage increases to 41% for Drosophila melanogaster proteins. Thus, intrinsically structural protein disorder is probably a common occurrence in vivo. It is unclear why structural disorder would confer a physiological advantage to the function of a protein function. Several possible reasons have been proposed to answer this question [39–44]. For example, if a protein is highly flexible, its association with various targets of different molecular dimensions and binding surfaces would be facilitated as different conformations might be assumed. Indeed, EspB prob- ably associates with various biological molecules, e.g. EspA [20], EspD [17], a-catenin [18] and a1-antitrypsin [19]. The molecular weights, physicochemical proper- ties and functions of EspA, EspD, a-catenin and a1-antitrypsin differ significantly. It is probable that different areas of EspB bind different target molecules. However, while EHEC and EPEC invade a variety of animals with target molecules of varying amino acid sequences, EspB should still specifically recognize isoforms of the target molecules at the same binding surfaces. Therefore, the conformational flexibility of a virulence factor should provide a mechanism that enables bacteria to infect various host species via the same infection system. Interestingly, exogenously added IpaC, an EspB homolog from S. flexneri, enhanced the invasion activ- ity of this bacterium into host cells [63]. As discussed above, IpaC assumes an almost fully unstructured con- formation near physiological conditions in vitro [64]. Such a property may also facilitate the penetration of this molecule into host cells. Thus, the conformational flexibility of EspB may also be advantageous for effi- cient penetration into host cell membranes. This idea is consistent with the concept that partial unfolding may be required for the insertion of protein toxins into host membranes [65]. Experimental procedures Expression and purification of EspB The cDNA, encoding EspB, was amplified from an EHEC E. coli O157:H7 cosmid library (RIMD 0509890, Sakai strain) [66,67] by PCR and cloned into a pT7 vector (Novagen). The full-length espB gene was subcloned into the expression vector pET28a (Novagen, Madison, WI, D. Hamada et al. EspB is a natively partially folded protein FEBS Journal 272 (2005) 756–768 ª 2005 FEBS 763 USA). The recombinant EspB has 20 amino acids, MGSS HHHHHHSSGLVPRGSH, added at the N terminus of the original sequence. Recombinant EspB was expressed in E. coli BL21(DE3) transformed with the afore mentioned plasmid. Cultures of Luria–Bertani (LB) broth, supplemented with 50 lgÆmL )1 of kanamycin, were inoculated with colonies and grown over- night at 37 °C with shaking. Then, a portion of each culture was diluted 100-fold into 1 L of fresh LB medium and incu- bated at 37 °C with shaking. Protein expression was induced by the addition of isopropyl thio-b-d-galactoside (at a final concentration of 1 mm) when cultures reached an attenuance (D)of% 0.6 at 600 nm. For the expression of protein uni- formly labeled with 15 N, M9 medium supplemented with 15 NH 4 Cl (Nippon Sanso Co., Kanagawa, Japan) was used instead of LB medium. After 4 h of further shaking at 37 °C, the cells were harvested by centrifugation (10 min, 10 000 g,4°C) and placed on ice. Protein was expressed as both soluble and insoluble fractions when the cells were dis- rupted by 20 mm sodium phosphate, pH 7.0, containing 0.1% (v ⁄ v) Triton X-100. However, the solubility was quite low. On the other hand, EspB cannot be extracted into the soluble fraction when the cells are disrupted by 20 mm sodium phosphate, pH 7.0. The purifications were therefore performed from either the soluble fraction obtained by dis- ruption of the cells in the presence of Triton X-100 or from the insoluble fraction obtained by cell disruption in the absence of detergent. For preparation from the soluble frac- tion, the cells were suspended with 20 mm sodium phos- phate, pH 7.0, containing 0.1% (v ⁄ v) Triton X-100 and the solution was separated by centrifugation (15 000 g, 10 min, 4 °C). The solution was loaded onto Chelating Sepharose Fast Flow (Amersham Biosciences, Corp., Piscataway, NJ, USA) supplemented with NiCl 2 in 20 mm sodium phos- phate, pH 7.0, washed with the same buffer and eluted using a 0–1.0 m imidazole gradient. The eluted protein was further purified with size-exclusion chromatography (S-300; 20 mm sodium phosphate, pH 7.0; Amersham Biosciences, Corp.). For preparation from the insoluble fraction, the cells were suspended in 20 mm sodium phosphate, pH 7.0, and lysed by sonication. The solution was centrifuged (15 000 g, 10 min, 4 °C) to separate the soluble and pellet fractions. The protein was extracted from the pellet by the addition of 20 mm sodium phosphate, pH 7.0, containing 8.0 m urea. This solution was clarified by centrifugation and diluted 100-fold by dropwise addition into 20 mm sodium phos- phate, pH 7.0, at 4 °C. The solution was then purified by Chelating Sepharose Fast Flow supplemented with NiCl 2 and further purified by size-exclusion chromatography (S-300) as in the case of preparation from the soluble frac- tions. The purification yields from soluble and insoluble fractions were 15 and 30 mg from 1 L of culture in LB medium, respectively. As judged by SDS ⁄ PAGE, the purity of recombinant EspB prepared from the insoluble fraction is relatively higher than that of EspB purified from the soluble fraction. According to the CD spectrum, both purifications yielded the same conformational state of EspB (see text for details). Owing to the higher yields of purification, 15 N pro- tein was prepared from insoluble fractions. The protein concentration was determined by absorption at 276 nm with the extinction coefficient of 4350 mÆcm )1 calculated from amino acid composition [68]. The protein solution was stored at )20 °C. CD spectroscopy CD spectra were measured by using a J-600 spectropola- rimeter (Jasco, Tokyo, Japan). The temperature was held at 20 °C by using a thermostatically controlled cell holder in conjunction with a circulating waterbath. For far-UV and near-UV CD spectra, cells of 1 mm and 1 cm path length were used, respectively. Protein concentrations were 0.1 and 1mgÆmL )1 for far-UV and near-UV CD measurements, respectively. The data are expressed as molar residue ellip- ticity [69], [h], with [h] ¼ 100 h obs (cl) )1 . The value, h obs ,is the observed intensity, c is the concentration in residue moles per litre, and l is the path length in cm. A secondary structure prediction was made by using the amino acid sequence of EspB in conjunction with the program package cdpro, in which selcon3, cdsstr and continll programs are included [70,71]. The reported values are the average of results from the above three programs. Fluorescence spectroscopy The fluorescence spectra of intrinsic EspB tyrosines and ANS were measured by using a FP-777 fluorimeter (Jasco). For tyrosine fluorescence, the excitation wavelength was 280 nm and the fluorescence emission was 300–350 nm. The protein concentration was 0.1 mgÆmL )1 . For ANS fluores- cence, the excitation wavelength was 350 nm and the emis- sion was measured between 400 and 650 nm. The protein concentration was 0.1 mgÆmL )1 and the ANS concentration was 5 lm. The temperature was maintained at 20 °C with a peltier-type thermostatically controlled cell holder. Fluorescence quenching of EspB tyrosines was measured in the presence of various concentrations of acrylamide, with spectra acquired as described above. Urea-induced unfolding measurements Urea-unfolding curves were plotted with [h] at 222 nm vs. the urea concentration. The data were analyzed assuming a two-state unfolding mechanism and assuming that the change in free energy of unfolding (DG), is linearly depend- ent on urea concentration: DG ¼ DG water À m½ureað2Þ Here, DG water corresponds to DG of unfolding in the absence of urea; m is a measure of the co-operativity EspB is a natively partially folded protein D. Hamada et al. 764 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS of the unfolding transition; and [urea] is the urea concen- tration. The fractions of unfolded (f U ) and folded (f F ) species at various urea concentrations can be expressed as: f U ¼ 1=½1 þ expðÀDGR À1 T À1 Þ ð3Þ and as: f F ¼ 1 À f U ð4Þ where R is the gas constant and T is the temperature in Kelvin. The theoretical value of [h] at 222 nm ([h] 222 ), observed in the presence of various concentrations of urea, can be expressed as: ½h 222 ¼½h F f F þ½h U f U ð5Þ Here, [h ] F and [h] U are the [h] 222 of the folded and unfolded species, respectively. The values for DG water and m were obtained by nonlinear curve fitting to the transition curves, according to Eqns (2–5), by using the program igorpro (WaveMetrics Inc., Lake Oswego, OR, USA). The linear dependences of [h] F and [h] U on urea concentrations were also considered in the fitting analysis. The same baselines for folded and unfolded species were used for the fitting of data obtained at pH 2.0–7.0. Multiangle laser light scattering Multiangle laser light scattering data were obtained by using a dynapro Molecular Sizing Instrument (Protein Solutions Inc., Milton Keynes, UK) at 20 °C. Various con- centrations of protein solution at pH 2.0, 4.0, 6.0 and 7.0 (400 lL) were passed through 0.22 lm of centrifugal filter unit, ultrafree-MC from Millipore (Billerica, MA, USA), and further centrifuged at 20 000 g for 10 min. Only the clear solution at the top of a tube (100 lL) was used for the light scattering analysis. Ultracentrifugation Sedimentation equilibrium experiments were performed by using a Beckman Optima XL-I analytical ultracentrifuge (Fullerton, CA, USA) at 11 300 g,20°C. The protein con- centration was 3 mgÆmL )1 . NMR spectroscopy 2D 15 N- 1 H HSQC spectra were recorded at 15 °C on either a 500 or an 800 MHz spectrometer (Brucker DRX500 or DRX800, respectively, Brucker Biospin GmbH, Karlsruhe, Germany), each equipped with a triple axis gradient and a triple-resonance probe. Protein concentrations were 1–2 mm in buffered solution containing 10% 2 H 2 O. For DRX500 experiments, the number of complex points and spectral widths were 1024, 12019 Hz ( 1 H, F 2 ) and 64, 1168 Hz ( 15 N, F 1 ), and for those using the DRX800 spectrometer, the parameters were 1024, 12821 Hz ( 1 H, F 2 ) and 64, 1866 Hz ( 15 N, F 1 ). The 1 H carrier was set at 4.7 p.p.m., and the 15 N carrier at 120 p.p.m. The 15 N- 1 H HSQC experiments included the WATERGATE and Water-flip- back techniques. The data were processed by using nmrpipe [72] and visualized by using sparky (TD Goddard & DG Kneller, University of California, San Francisco, CA, USA; http://www.cgl.ucsf.edu/home/sparky/). Acknowledgements The authors acknowledge Prof. Yuji Goto for access to the CD spectropolarimeter, Prof. Atsushi Nakagawa for access to light scattering, and Miyo Sakai for per- forming ultracentrifugation. This work was supported, in part, by grants-in-aid for scientific research from the Japan Ministry of Education, Science, Culture and Sports, and JSPS Research Fellowships for Young Sci- entists (to D.H.). References 1 Galan JE & Collmer A (1999) Type III secretion machines: bacterial devices for protein delivery into host cells. Science 284, 1322–1328. 2 McDaniel TK, Jarvis KG, Donnenberg MS & Kaper JB (1995) A genetic locus of enterocyte effacement con- served among diverse enterobactrial pathogens. Proc Natl Acad Sci USA 92, 1664–1668. 3 McDaniel TK & Kaper JB (1997) A cloned pathogeni- city island from enteropathogenic Escherichia coli con- fers the attaching and effacing phenotype on K-12 E. coli. Mol Microbiol 23, 399–407. 4 Perna NT, Mayhew GF, Po ´ sfai G, Eliott SJ, Donnenberg MS, Kaper JB & Blattner FR (1998) Molecular evolution of a pathogenicity island from enterohaemorrhagic Escherichia coli O157:H7. Infect Immun 66, 3810–3817. 5 Zhu C, Agin TS, Elliott SJ, Johnson LA, Thate TE, Kaper JB & Boedeker EC (2001) Complete nucleotide sequence and analysis of the locus of enterocyte efface- ment from rabbit diarrheagenic Escherichia coli RDEC-1. Infect Immun 69 , 2107–2115. 6 Shuch R & Maurelli AT (2000) The Type III secretion pathway. Dictating the outcome of bacterial–host inter- actions. In Virulence Mechanisms of Bacterial Pathogens, 3rd edn (Brogden KA, Roth JA, Stanton TB, Bolin CA, Minion FC & Wannemuehler MJ, eds), pp. 203– 233. ASM Press, American Society for Microbiology, Washington, DC, USA. 7 Nougayre ` de J-P, Fernandes PJ & Donnenberg MS (2003) Adhesion of enteropathogenic Escherichia coli to host cells. Cell Microbiol 5, 359–372. D. Hamada et al. EspB is a natively partially folded protein FEBS Journal 272 (2005) 756–768 ª 2005 FEBS 765 [...]... Kurokawa K, Ishii K, Yokoyama K, Han CG, Ohtsubo E, Nakayama K, Murata T et al (2001) Complete genome sequence of enterohaemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12 DNA Res 8, 11–22 67 Perna NT, Plunkett G, III, Burland V, Mau B, Glasner JD, Rose DJ, Mayhew GF, Evans PS, Gregor J, Kirkpatrick HA et al (2001) Genome sequence of 767 EspB is a natively partially. . .EspB is a natively partially folded protein 8 Kenny B, DeVinney R, Stein M, Reinscheid DJ, Frey EA & Finlay BB (1997) Enteropathogenic E coli (EPEC) transfers its receptor for intimate adherence into mammalian cells Cell 91, 511–520 9 Vlademir VC, Takahashi A, Yanagihara I, Akeda Y, Imura K, Kodama T, Kono G, Sato Y & Honda T (2001) Talin, a host cell protein, interacts directly with the translocated... Tir, of enteropathogenic Escherichia coli, and is essential for pedestal formation Cell Microbiol 3, 745–751 10 Vlademir VC, Takahashi A, Yanagihara I, Akeda Y, Imura K, Kodama T, Kono G, Sato Y, Iida T & Honda T (2002) Cortactin is necessary for F-actin accumulation in pedestal structure induced by enteropathogenic Escherichia coli infection Infect Immun 70, 2206–2209 11 Crane JK, McNamara BP & Donnenberg... 3, 669–679 18 Kodama T, Akeda Y, Kono G, Takahashi A, Imura K, Iida T & Honda T (2002) The EspB protein of enterohaemorrhagic Escherichia coli interacts directly with a- catenin Cell Microbiol 4, 213–222 19 Knappstein S, Ide T, Schmidt MA & Heusipp G (2004) a1 -Antitrypsin binds to and interferes with functionality of EspB from atypical and typical enteropathogenic Escherichia coli strains Infect Immun... partially folded protein enterohaemorrhagic Escherichia coli O157:H7 Nature 409, 529–533 68 Gill SC & von Hippel PH (1989) Calculation of protein extinction coefficients from amino acid sequence data Anal Biochem 182, 319–326 69 Hamada D, Kuroda Y, Kataoka M, Aimoto S, Yoshimura T & Goto Y (1996) Role of heme axial ligands in the conformational stability of the native and molten globule states of horse... Hagihara Y, Tan Y & Goto Y (1994) Comparison of the conformational stability of the molten globule and native states of horse cytochrome c J Mol Biol 237, 336–348 FEBS Journal 272 (2005) 756–768 ª 2005 FEBS EspB is a natively partially folded protein 53 Hamada D, Kidokoro S, Fukada H, Takahashi K & Goto Y (1994) Salt-induced formation of the molten globule state of cytochrome c studied by isothermal... of EspA filaments from enteropathogenic Escherichia coli Mol Microbiol 49, 301–308 Sekiya K, Ohishi M, Ogino T, Tamano K, Sasakawa C & Abe A (2001) Supermolecular structure of the enteropathogenic Escherichia coli type III secretion system and its direct interaction with the EspA-sheath-like structure Proc Natl Acad Sci USA 98, 11638–11643 Knutton S, Rosenshine I, Pallen MJ, Nisan I, Neves BC, Bain C,... Sreerama N & Woody RW (2000) Estimation of protein secondary structure from circular dichroism spectra: 768 D Hamada et al comparison of CONTIN, SELCON, and CDSSTR methods with an expanded reference set Anal Biochem 287, 252–260 71 Sreerama N, Venyaminov SY & Woody RW (2000) Estimation of protein secondary structure from circular dichroism spectra: inclusion of denatured proteins with native proteins... 65–78 16 Roe AJ, Hoey DE & Gally DL (2003) Regulation, secretion and activity of type III-secreted proteins of enterohaemorrhagic Escherichia coli O157 Biochem Soc Trans 31, 98–103 17 Ide T, Laarmann S, Greune L, Schillers H, Oberleithner H & Schmidt MA (2001) Characterization of translocation pores inserted into plasma membranes by type IIIsecreted Esp proteins of enteropathogenic Escherichia coli Cell... isothermal titration calorimetry Proc Natl Acad Sci USA 91, 10325–10329 54 Bertagna AM & Barrick D (2004) Nonspecific hydrophobic interactions stabilize an equilibrium interemediate of apomyoglobin at a key position within the AGH region Proc Natl Acad Sci USA 101, 12514– 12519 55 Marmorino JL, Lehti M & Pielak G (1998) Native tertiary structure in an A- state J Mol Biol 275, 379–388 56 Redfield C, Smith RA & Dobson . EspB from enterohaemorrhagic Escherichia coli is a natively partially folded protein Daizo Hamada 1 , Tomoaki Kato 1,2 , Takahisa Ikegami 3 , Kayo. the raw data. The line is drawn only for visual assistance and is not a mathematical fit. Fig. 5. Urea unfolding of EspB at various pH values and at 20