INTRODUCTION

Recent structural studies on proteins have revealed that some are constructed by "modules," which are compact structural units (1, 3). The exon shuffling in protein evolution with the correspondence of the module to the exon on the gene structure strongly suggested that the module structure is one of the key factors in evolution of structure and function of protein (4, 5). The module structure was first identified in the globin family (1), and their diagonal plots have clearly shown that both of the Hb subunits (- and -subunits) consist of the four modular structures, the modules M1, M2, M3, and M4 (1), each of which is coded by an exon with the exception of the large central exon comprising two compact units.

The analysis of hemoglobin functions showed that the amino acid residues associated with the heme contacts are concentrated in the central modules (M2 + M3) as illustrated in Fig. 1 (6). In 1980, by using a protease, Craik et al. (7, 8) isolated the central region of the -subunit, which corresponds to the module M2 + M3 and is coded by the central exon of the -subunit. They showed that the fragment from the central region bound heme stoichiometrically and tightly, generating a characteristic strong Soret absorption band as found for the intact subunits. De Sanctis et al. (9-12) also prepared "minimyoglobin," which is constructed by the peptide fragment from 32 (Leu) to 139 (Asn) of horse heart myoglobin. The heme binding property of the minimyoglobin peptide was almost the same as that of the intact myoglobin, and the heme binding induced the recovery of secondary structure of the minimyoglobin. They concluded that the removal of the peptide fragments from the amino and carboxyl terminus has little effect on the heme binding to the apoprotein. These experimental results from the limited proteolysis of globins indicate that the module M2 + M3 can be a structural and functional unit to bind heme.

[View Larger Version of this Image (26K GIF file)]

To obtain further insight into the modular structure in globins, we have focused on the module substitution by use of cassette mutagenesis (2). Our previous results have revealed that the heme environmental structure and ligand binding property of the chimeric -subunit, in which module M4 of the -subunit was replaced by that of the -subunit, were quite similar to those of the -subunit, while the chimeric subunit preferentially bound to the native -subunit to form a heterotetramer, ()₂()₂, not to the native -subunit. The conversion of the association property indicates that module M4 is a key module in specific binding of the -subunit to the -subunit. This conclusion also corresponds to the distribution of the amino acid residues responsible for the formation of the 1-1-subunit interface as shown in Fig. 1.

In the present study, we prepared the "reversed" chimeric subunit, the -subunit, in which module M4 of the -subunit is replaced by that of the -subunit, to clarify structural and functional roles of the module M4 in more detail. We examined the contributions of the modular structure to the stability of globular proteins and their functional roles in regulating the association property by comparing structural and functional properties of the chimeric - and -subunits.

EXPERIMENTAL PROCEDURES

To construct the mutated gene for the chimeric -subunit, the phage vector M13mp18 containing the -subunit gene was digested with HindIII and NdeI (13). Since the native -subunit gene has no HindIII site at the corresponding position, a HindIII site was introduced at position FG3 in the -subunit gene on M13mp18 cII FX -subunit (14-16) without changing the encoded amino acid residue (2). The HindIII-NdeI fragment of the -subunit and the HindIII-HindIII fragment of the mutant -subunit were purified on agarose gel. The fragments were ligated into the HindIII-NdeI fragment of M13mp18 to form M13mp18 cII FX chimeric -subunit.¹ The ligated fragment was then cloned into the HindIII-NdeI fragment of the T7 expression vector.

The prepared gene encoding the chimeric subunits were transformed into an Escherichia coli strain (BL21), which was grown at 37 °C in 2xTY culture containing ampicillin (100 µg/ml) overnight. This culture was centrifuged at 5,000 rpm, and the cell pellets were lysed with lysozyme (100 mg/10 g of pellets) to get the crude chimeric subunits. By purification processes with gel chromatography (Sephacryl-200) and ion-exchanged column (CL-6B), we obtained chimeric subunit purity of more than 90% as estimated on SDS gels. Fast atom bombardment mass spectroscopy on trypsin-digested proteins (17) showed no unexpected mutations in the chimeric subunits (data not shown).

Heme titration was monitored by the absorbance change at 413 nm (Hitachi U-3210). To estimate the concentration of the chimeric subunits, we used the extinction coefficient at 280 nm, which was determined to be 8.8 mM¹ cm¹ for the chimeric -subunits and 13.2 mM¹ cm¹ for the chimeric -subunits by the number of the aromatic residues, tryptophan (₂₈₀ = 5.5 mM¹ cm¹) and tyrosine (₂₈₀ = 1.1 mM¹ cm¹) (18). The carbon monoxide derivative was prepared by adding dithionite to the ferric chimeric subunits under CO atmosphere. No endogenous ligand-bound derivative was gained with recourse to completely substituting N₂ for CO in the surrounding atmosphere. Absorption spectra were recorded with a Hitachi U-3210 spectrophotometer.

CD spectra of the chimeric and native subunits in the far-UV and Soret regions were measured at 20 °C with a Jasco J-760 spectrometer. The concentration of the sample was 10 µM, and the buffer was 50 mM Tris, 0.1 M NaCl, 10 mM NaCN, and 1 mM Na₂EDTA, pH 7.4. The path lengths for the measurements in the far-UV and Soret regions were 1 and 10 mm, respectively. The mean residue -helical contents were estimated with the deconvolution method by Greenfield and Fasman (19), where the chimeric globin subunits are assumed to contain only -helical and random coil structure, since the native globin subunits have no -strand.

Tryptophan fluorescence spectra were measured with a Perkin-Elmer LB50 emission spectrometer. The samples were excited at 280 nm, and the emission spectra were measured between 300 and 400 nm. We used a crystal cuvette for the fluorescence measurements, and the path length was 10 mm. The concentration of the samples was 5 µM on a heme basis, and the buffer was the same as used in the CD measurements.

Reaction solutions contained 50 mM sodium phosphate (pH 7.4), 1 mM NaCN, and various concentrations of urea. Changes in ellipticity at 222 nm, []₂₂₂, were monitored with a Jasco J-760 CD spectrophotometer after several hours of equilibration at room temperature. The fractional population of denatured form (f_D) under each condition was estimated by the equation,

(Eq. 1)

where []_222, N and []_222, D represent ellipticities at 222 nm in the native and denatured state, respectively (20).

Solution x-ray scattering experiments on native Hb and the chimeric globin subunits were carried out at the solution scattering station (SAXS camera) installed at BL-10C (the Photon Factory, Tsukuba, Japan) (21, 22). The sample-to-detector distance was about 90 cm for measurements of SAXS, calibrated by meridional diffraction of dried chicken collagen. The wavelength of the x-ray was 1.488 Å. The sample cell was 50 µl in volume with 15-µm-thick quartz windows and had a 1-mm x-ray path length. The temperature of the sample was controlled at 25 °C by circulating the temperature-controlled water. The protein concentrations were varied within a range of 3-10 mg/ml, and the measurement time for each scattering experiment was 10 min. The scattering from the solution containing no protein was measured to subtract background scattering.

Data processing was carried out with Apple Macintosh personal computers. The scattering curve at infinite dilution was obtained from a series of scattering data with different protein concentrations (23). X-ray scattering intensities at the small angle region are given as the equation,

(Eq. 2)

where Q and I(0) are the momentum transfer and the intensity at 0 scattering angle, respectively (24). Q is defined by Q = 4sin/, where 2 and  are the scattering angle and the wavelength of the x-ray, respectively. The radius of gyration (Rg) was obtained from the slope of the Guinier plot, a plot of ln(I(Q)) against Q² (25).

Measurements were performed by gel filtration on a Sephacryl S-200 HR column (0.8 × 62 cm) at 4 °C. The buffer used for the chromatography was 50 mM Tris, 0.1 M NaCl, and 1 mM Na₂EDTA, pH 7.4, and the flow rate was 5 ml/h. Sample concentration was 15 µM, and eluted fractions were monitored at the Soret absorption band.

RESULTS

The spectrophotometric titration of heme to the chimeric - and -subunits showed that the chimeric subunits bound heme stoichiometrically (data not shown). The absorption spectra of the carbonmonoxy chimeric subunits were identical to that of the native subunit as illustrated in Fig. 2A. However, the deoxygenated chimeric -subunit exhibited a quite different spectrum from that of deoxy native subunit (Fig. 2B), while the chimeric -subunit showed the deoxy-type spectrum as found for native subunit (2). In the deoxygenated form, the absorption maxima were observed at 531 and 559 nm in the Q band of the chimeric -subunit, whereas native subunit (-subunit) and the chimeric -subunit have a single and broad peak at 555 nm. The peak position of the Soret band also showed a blue shift for the deoxy chimeric -subunit. This spectral pattern is characteristic of the hemochromogen-type hemoprotein, in which the distal histidine as well as the proximal histidine is coordinated to iron(II) (26).

[View Larger Version of this Image (13K GIF file)]

In Fig. 3A, the far-UV CD spectra of the cyano-met native and chimeric subunits are shown. The native - and -subunits exhibited two negative broad peaks around 222 and 208 nm (27). In the spectra of the chimeric subunits, a remarkable decrease in the negative ellipticity around 222 nm was observed, and a new negative peak around 206 nm appeared. Based on the deconvolution method by Greenfield and Fasman (19), the -helical content of the chimeric -subunit was estimated as 30%, which is much lower than that of the native subunits (70%) and the chimeric -subunit (68%). In the Soret region, the CD spectrum of hemoprotein has also provided some information of the heme environmental structure (28). As shown in Fig. 3B, the CD spectra in Soret region for cyano-met native and chimeric -subunits exhibit maxima around 420 nm, which is the characteristic Cotton effect in hemoproteins (28). In sharp contrast to the large Cotton effect in native and chimeric -subunits, the CD spectra of the -subunit shows a small maximum at 417 nm, suggesting that the substitution of module M4 induced large structural changes in the heme environment of the chimeric -subunit but not so much of the -subunit.

[View Larger Version of this Image (19K GIF file)]

Fig. 4 illustrates the fluorescence spectra of the -, -, chimeric -, and chimeric -subunits. The -subunit showed a small peak around 310 nm originating from tyrosines (Tyr²⁴, Tyr⁴², and Tyr¹⁴⁰) (29), while the fluorescence peak of the -subunit around 335 nm is characteristic of Trp³⁷ (30). In the chimeric proteins, the -subunit indicated a large peak around 345 nm, in addition to the peak at 310 nm, although the chimeric subunit contains only one tryptophan (Trp¹⁴) in the A-helix. In the spectra of the chimeric -subunit, a broad peak appeared around 335 nm with a shoulder at 310 nm as observed for the -subunit, but the intensity was greater than that of the -subunit.

[View Larger Version of this Image (17K GIF file)]

The structural stability of the chimeric - and -subunits was examined by urea denaturation.² Urea denaturation curves for Hb tetramer, native -, -, chimeric -, and chimeric -subunits are shown in Fig. 5. As clearly shown in the Fig. 5, the transition curve for the chimeric -subunit largely shifted to the left side and did not show cooperativity, which implies large destabilization caused by the substitution of module M4. On the other hand, the denaturation curve for the chimeric -subunit was cooperative and similar to those of native subunits, indicating that structural stability is not perturbed by the module substitution in the chimeric -subunit.

[View Larger Version of this Image (22K GIF file)]

Fig. 6A shows Guinier plots of SAXS data for native Hb and the chimeric - and -subunits. The curves for Hb and the chimeric -subunit are well approximated by a straight line in the small angle region (the Guinier region) (25). Their Rg values were deduced from the slopes of the regression lines within the Guinier region, defined as Rg·Q < 1.3 in the present study (31). The Rg value for the chimeric -subunit is 25.9 Å, which is slightly larger than that for Hb tetramer (25.3 Å). By the linear fitting analysis, the almost identical values of the intensity at 0 scattering angle, I(0), was obtained for Hb tetramer (2.18 ± 0.03 arbitrary units) and the chimeric -subunit (2.18 ± 0.06 arbitrary unit). Assuming the same specific volumes for Hb tetramer and the chimeric -subunit, these results for the I(0) values suggest the same molecular mass of Hb and the chimeric -subunit. On the other hand, some oligomerizations were observed for the chimeric -subunit solutions. Although the accurate Rg and I(0) values of the -subunit could not be determined due to the deviation from the regression line, the I(0) value can be larger than those of Hb and the chimeric -subunit (Fig. 6B), probably due to aggregation and/or unstable tertiary fold.

[View Larger Version of this Image (14K GIF file)]

The Kratky plots of Hb and the chimeric subunits are shown in Fig. 6C. Hb indicated a clear peak fitted to a quadratic function, which is typical for a rigidly folded globular protein. The plot pattern for the chimeric -subunit is almost identical to that of Hb, implying that the global conformation of the chimeric -subunit is essentially the same as that of Hb. For the chimeric -subunit, however, the peak intensity in the Kratky plot was greatly decreased. The smaller descent, or larger slope, toward larger values of Q suggests a partly disordered structure of the chimeric -subunit, since this plot pattern is clearly different from that of completely denatured proteins (24).

Gel chromatograms of the carbonmonoxy chimeric -subunit in the presence and absence of native subunits are shown in Fig. 7. Under the conditions applied here, the mixture of native - and -subunits formed tetramers, whereas the -subunit still remained in a monomer (32). The -subunit was in the equilibrium between monomers and tetramers (32). As shown in Fig. 7, the elution peak for the chimeric -subunit was observed at the position for a tetrameric globin, which indicates that the chimeric -subunit forms the homotetrameric protein, ()₄, by self-association. In the chromatogram of the mixture of the chimeric - and -subunits, two peaks were observed, which coincided with the peaks for the isolated chimeric - and -subunit. On the other hand, the peak for the mixture of the chimeric - and -subunits was twice as high as that for the isolated chimeric -subunit, implying that the chimeric -subunit associates with the -subunit to form a heterotetramer, not with the -subunit. These association properties were independent of the sample concentration from 1 to 50 µM.

[View Larger Version of this Image (21K GIF file)]

We also examined salt effects on the association property of the chimeric globins by gel chromatography in the presence of NaCl (up to 1 M). Previous studies have shown that the addition of NaCl drastically increases dissociation of Hb tetramer, whereas the tetrameric -subunit (₄) is relatively insensitive to salt or slightly stabilized in salt. These different responses to the addition of salt were attributed to the different intersubunit interactions between the two tetrameric globins (32, 33). The elution pattern of the chimeric -subunit was almost insensitive to NaCl (data not shown), suggesting that the subunit interactions in the -subunit are different from those of both of the native tetrameric globins.

DISCUSSION

Although the heme titration clearly showed that the chimeric subunits stoichiometrically bound heme as found for the native subunit, the hemochromogen-type spectrum was observed for the deoxygenated chimeric -subunit (Fig. 2). Since the hemochromogen-type spectrum is one of the characteristics for the denatured or partially unfolded hemoproteins (26), the heme environmental structure of the chimeric -subunit was significantly perturbed upon the substitution of module M4. The perturbation to the globin structure by the module substitution is also manifested in the CD spectra as illustrated in Fig. 3. The decreased Cotton effect in the Soret region for the chimeric -subunit can be interpreted as large structural changes in its heme vicinity, and the small negative ellipticity around 222 nm in the chimeric indicates that the module substitution also induced global structural changes in the chimeric protein.³ Since the mean residue ellipticity at 222 nm depends on the dynamic motion of the protein and the precise conformation of helical residues (34), the lower helical content implies that the structure of the chimeric -subunit would not be a "rigid" globular conformation as found in native globins and that the fragmentation of long helices might be induced by the module substitution of module M4 in the -subunit.

Tryptophan fluorescence spectra also strengthen the suggestion of a partially unfolded structure of the chimeric -subunit. In the native -subunit, since Trp¹⁴ is fully buried near the center of the molecule, the fluorescence of Trp¹⁴ is quenched due to resonance energy transfer to the heme moiety and low intensity (30). Upon substitution of module M4, the remarkable increase in the intensity of the fluorescence peak around 345 nm is observed in the chimeric -subunit, suggesting drastic changes in the microenvironment of Trp¹⁴ and heme moiety. The large peak at 345 nm implies that Trp¹⁴ is located in a hydrophilic and exposed environment (35), which leads us to speculate that the protein structure around Trp¹⁴ does not completely fold in the chimeric -subunit as found in denatured proteins (35).

A partially unfolded structure in the chimeric -subunit is confirmed by the SAXS measurements. A large I(0) value in the Guinier plot for the chimeric -subunit suggests aggregation, corresponding to many fluctuations and/or partial exposure of hydrophobic residues in the chimeric -subunit. In the Kratky plot, the peak for the chimeric -subunit is low and asymmetric, which is also characteristic of a partially denatured globular structure (24). Together with these spectral data, the chimeric -subunit does not fold as does the native subunit, and some parts of the chimeric protein appear to remain unfolded and largely fluctuate. Such large fluctuation would be related to the destabilization of the chimeric -subunit as suggested in the urea denaturation experiment, where melting of the secondary structure was observed in the chimeric -subunit at a urea concentration lower than in the native -subunit. The noncooperative denaturation of the chimeric -subunit also implies that its folded structure is different from the - and -subunits of native globin.

On the other hand, as our previous studies have shown, the structure of the -subunit is quite similar to that of the -subunit (2). In the present study, the spectral features in the CD spectra of the chimeric -subunit are typical of the globin proteins, indicating that the heme environment and secondary structure are retained despite this module substitution. Also, based on the fluorescence spectra, Trp³⁷ and Trp¹⁵ of the chimeric -subunit are located in a similar environment to those of the -subunit. The stable globin structure in the -subunit is manifested in the SAXS and urea denaturation measurements. The Rg and I(0) values for the chimeric -subunit were estimated to be the same as those of native Hb tetramer. The denaturation curve for the chimeric -subunit is similar to that of Hb.

It is quite interesting that the chimeric - and -subunits exhibited remarkable differences in their structure and stability. Based on the computer modeling of the -subunit, the side chains of Tyr¹²⁵ (in module M4 in the -subunit) would be located near the indole group of Trp¹⁴ (in module M1 of the -subunit), which might induce some steric constraint between the two side chains. Since Jennings and Wright (36) showed that the packing between the A (M1), G (M4), and H helices (M4) is formed in the first step of the folding process in apomyoglobin and since these helices are crucial in protein folding of globins, the steric hindrance between the side chains of Tyr¹²⁵ (M4) and Trp¹⁴ (M1) could interfere with the packing of modules M1 and M4, thereby resulting in the destabilization of the globular structure of the chimeric -subunit. The fluorescence spectra for the chimeric -subunit also indicate that the microenvironment around Trp¹⁴ is very different from that in the native -subunit. Such steric hindrance cannot be found between modules M1 and M4 of the chimeric -subunit, which leads to the formation of the stable packing between the helices of modules M1 and M4. Although the quantitative analysis and comparison of the effects of the module substitution on the globin structure between the two chimeric globins have not yet been done due to a lack of structural information, it can be safely said that there is less prominent steric hindrance in the -subunit than in the -subunit.

In a previous paper (2), we suggested that the modules are structural and functional units that have the advantage of producing stable functional proteins by their combinations. However, the large structural changes and destabilization for the chimeric -subunit, which were never observed in the chimeric -subunit, indicate that the simple substitutions of the modules are not enough to construct stable globins. The steric constraint between side chains exerted by the module substitution would prevent the chimeric globin from forming the stable globular structure as discussed above, suggesting that interactions between modules are essential to stabilize the globular structure.

A schematic representation for the subunit interactions in the mixture of the chimeric and native subunits is given in Fig. 8 (6). In the chimeric -globin, the module substitution would create some interactions between the modules M2 + M3 and M4 of the two chimeric -subunits and also between their modules M4, which are not found for the -subunits. These new subunit interactions formed by the substitution of module M4 would lead to the formation of homotetramer, ()₄. Although the elution peak for the chimeric -subunit appeared at the position of the tetrameric globin, the broad elution peak for the chimeric -subunit suggests the formation of oligomeric globins other than the tetrameric globin. Such a broad peak was also encountered for unstable protein solution, in which the protein has a high molecular dispersity (37). The Guinier plot for the chimeric -subunit shows a negative steep gradient in the small Q region, which is also suggestive of the existence of several oligomeric globins. We also cannot rule out the possibility of elongated dimer formation in the chimeric -subunit. Previous studies on the molten-globule state of apomyoglobin and cytochrome c pointed out that the molecular radius of a partly unfolded dimer state could be enlarged as much as that of a stable tetramer in the native state (38, 39), so that it is possible that the elution volume of the chimeric -subunit coincides with that of native tetrameric Hb in the gel chromatogram.

Interactions between the modules of the chimeric -subunits and those of the native subunits.

[View Larger Version of this Image (42K GIF file)]

The unstable structure in the associated chimeric -subunit can be interpreted in terms of the differences of the subunit interfaces between the chimeric and native subunits. As the CD and fluorescence spectra show, it is unlikely that the structure of the interface in the tetrameric chimeric globins was conserved as found for the native subunit, which is also supported by the effects of NaCl on the association property. In the native Hb tetramer, the addition of NaCl facilitates its dissociation into dimers, due to perturbation of the salt bridges and hydrogen bonds, which play a crucial role in 1-2-subunit interface (40). The negligible effect of NaCl on the association property of the chimeric -subunit suggests that electrostatic interactions such as hydrogen bonds and salt bridges would not be essential in the interface of the associated chimeric globin. Based on the remarkable dispersity of the associated states for the chimeric -subunit (Figs. 6 and 7), the structural disorder would induce unspecified interactions between the exposed hydrophobic residues, which might be responsible for the association in the chimeric -subunit.

In the presence of the native -subunit, the elution pattern indicated that the homotetramer of the chimeric -subunits dissociates and that an ()₂₂ heterotetramer forms. Since the crystallographic study of the native -subunit showed that module M4 of the -subunits interacts with module M4 of the other -subunits (33, 41, 42), it is plausible that there are many interactions between the chimeric - and the native -subunit to form the stable heterotetramer as shown in Fig. 8.

On the other hand, the addition of the -subunit did not induce the dissociation of the tetrameric -subunit, and the formation of the heterotetramer was never observed. Due to weak interactions between the two -subunits (32), interactions for formation of the heterotetramer between the native - and chimeric -subunits are only formed between module M4 of the chimeric -subunit and modules M2 + M3 and M4 of the -subunit (Fig. 8), which would be much weaker than those of the homotetramer, ()₄.

In summary, the substitution of module M4 in the -subunit for that of the -subunit induces structural alterations in the heme vicinity as well as in the global structure of globins. The resulting structural defects would originate from the failure of the stable packing of the helices in modules M1 and M4. A comparison with our previous results for the chimeric -subunit suggests that the module cannot be a structural unit without the stable and proper interactions between the modules. Thus, we can conclude that, although module substitution would be useful in the first step to design novel functional proteins, structural collapse must be considered in module-based protein design, and fine tuning by site-directed or random mutagenesis is required to recover protein structure and optimize function.