J Biol Chem, Vol. 274, Issue 43, 30357-30360, October 22, 1999

COMMUNICATION
UV Resonance Raman Spectra of Ligand Binding Intermediates of Sol-Gel Encapsulated Hemoglobin^*,

Laura J. Juszczak and Joel M. Friedman

From the Department of Physiology and Biophysics, Albert Einstein College of Medicine, Bronx, New York 10461

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

We report for the first time specific conformational changes for a homogeneous population of ligand-bound adult deoxy human hemoglobin A (HbA) generated by introducing CO into a sample of deoxy-HbA with the effector, inositol hexaphosphate, encapsulated in a porous sol-gel. The preparation of ligand-bound deoxy-HbA results from the speed of ligand diffusion relative to globin conformational dynamics within the sol-gel (1). The ultraviolet resonance Raman (UVRR) difference spectra obtained reveal that E helix motion is initiated upon ligand binding, as signaled by the appearance of an 1415 Trp W3 band difference at 1559 cm¹. The subsequent appearance of Tyr (Y8a and Y9a) and W3 (1549 cm¹) UVRR difference bands suggest conformational shifts for the penultimate Tyr140 on the F helix, the "switch" region Tyr42, and the "hinge" region Trp37. The UVRR results expose a sequence of conformational steps leading up to the ligation-induced T to R quaternary structure transition as opposed to a single, concerted switch. More generally, this report demonstrates that sol-gel encapsulation of proteins can be used to study a sequence of specific conformational events triggered by substrate binding because the traditional limitation of substrate diffusion times is overcome.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

A molecular level understanding of how proteins function requires not only a three-dimensional picture of equilibrium structures but also a time-ordered sequence of functionally related molecular events. The second requirement is difficult because of the challenge of either trapping detectable populations of nonequilibrium structures or creating a synchronized nonequilibrium population whose dynamics can be followed.

Because protein dynamics can span many decades in time starting on the picosecond time scale, it is often necessary to prepare the evolving population using a short, laser-derived excitation pulse. This approach has been very useful when starting with ligand-saturated hemeproteins. Nanosecond and faster photodissociation of ligand-bound hemoglobins has shed considerable light on the sequence of molecular steps associated with the tertiary and quaternary structure changes in the globin that follow ligand dissociation (2-10).

The sequence of conformational events following solvent-derived (non-geminate) ligand binding to deoxy hemoglobin is much more of a challenge because of the temporal constraints imposed by ligand diffusion. The initial conformational responses to ligand or substrate binding to a given protein will typically be faster than the diffusion time that is inherent to any rapid mixing experiments. In the present work, an approach to overcoming this limitation, based on sol-gel encapsulation, is presented in which the conformational response time is extended well beyond the ligand diffusion time. Thus we have been able to use UVRR¹ to probe the initial sequence of conformation changes in hemoglobin when CO binds to the equilibrium form of deoxy-HbA.

The encapsulation of proteins in porous TMOS-derived sol-gels has been shown to occur without significant alteration of protein structure and function (11, 12). Encapsulated proteins are in contact with the solvent through the approximately 50-100-Å diameter gel pores and are capable of binding or releasing small ligands or substrate molecules.

Sol-gel encapsulated HbA binds and releases oxygen non-cooperatively (13-15), implying that encapsulation limits ligand-induced quaternary structure changes. Furthermore, the results indicate that a decrease in temperature from ambient to below 10 °C enhances the capability of the sol-gel to limit ligand binding-induced quaternary structure changes. More recently (1), it has been directly shown via resonance Raman spectroscopy that encapsulation dramatically slowsbut not prevents-both the nanosecond to microsecond tertiary changes and the slower quaternary structure changes that follow the dithionite-mediated removal of oxygen from oxy-HbA. At 4 °C, minimal relaxation is observed out to 10³ s. At higher temperature, both tertiary and quaternary relaxation phases are observed, but on time scales that are orders of magnitude slower than in free solution at the same temperature. At 80 °C, the complete relaxation takes on the order of 2 h in contrast to 4 °C, where much of the relaxation is essentially frozen for days. This approach allows for probing the relaxation of a fully dissociated oxy-HbA population without the complications that occur with photodissociated samples in which geminate rebinding generates a mixed population of partially liganded species.

A key observation in that earlier study (1) is that, at lower temperatures, the addition of dithionite and its subsequent diffusion and oxygen scavenging can be made to occur faster than the relaxation times of hemoglobin. This finding stimulated the current study by raising the prospect of following the T to R transition in HbA by adding ligand to a cooled, encapsulated deoxy hemoglobin sample.

Quaternary structure dynamics in the sol-gel matrix have been previously studied by rapidly increasing the temperature of encapsulated samples of partially liganded iron-metal hybrid hemoglobin, and following kinetics using absorption spectroscopy (16). The absorbance changes seen in the broad Soret and visible bands during this transition provide no details about the critical movement of specific amino acid residues within the globin. In the present study, UVRR spectroscopy (17-19) is used to probe the ligand binding-induced conformational changes of tyrosine and tryptophan residues in HbA. Conformational changes are probed both in the critical "switch" and "hinge" regions of the R-T sensitive ₁₂ interface and in the scaffolding of the distal heme pocket, as reflected in the packing of the E and A helices. The ligand diffusion problem is overcome, as discussed above, by adding CO to TMOS-encapsulated deoxy-HbA under conditions where the T state is favored (high ionic strength and added effector, IHP) and conformational changes are highly constrained (6 °C).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

CW laser excitation at 229 nm was provided by an intracavity frequency-doubled Ar²⁺ laser. The power at the sample was 1.7 milliwatt. The UVRR system is described elsewhere (20). Adult human hemoglobin A was a generous gift of Dr. Seetharama Acharya. Concentrated HbA was diluted to 0.125 mM with 50 mM Hepes, at pH 7.5, for solution samples. Deoxygenation was carried out on nitrogen-purged samples in an anaerobic atmosphere via addition of a 4-fold excess (4× heme concentration) of sodium dithionite (J. T. Baker, Phillipsburg, NJ). CO ligation was accomplished by gaseous replacement. The hemoglobin ligation state was monitored by absorption spectroscopy of the visible and Soret bands (Lambda 2 UV/VIS spectrometer, Perkin-Elmer). 0.3 M perchlorate was added as an internal standard (sodium perchlorate, 0.002% chloride maximum, EM Science, Gibbstown, NJ). Sol-gels were prepared according to the protocol given in Refs. 11 and 12. Hemoglobin concentration in the sol is 0.125 mM with a 6-fold excess of the effector, IHP. Deoxy-HbA sol-gel "pipes," measuring 10 mm outer diameter × 9 mm inner diameter × 25.4 mm height, were cast in quartz NMR tubes using a Teflon mold under a nitrogen atmosphere. Subsequent to gelation, the deoxy-HbA samples were bathed in nitrogen-purged 50 mM Hepes, pH 7.5, buffer with the addition of 0.3 M perchlorate as an internal UVRR standard. The NMR tubes were subsequently sealed with rubber septa. The sol-gel samples were then wet-aged for 20 h at 4 °C under a nitrogen atmosphere. After acquisition of the UVRR spectrum of encapsulated deoxy-HbA, the bathing buffer solution was exchanged with CO-saturated buffer, and the cavity space above the sol-gel was exchanged with CO gas. The time of CO addition was recorded, as well as the time of UVRR spectral acquisition for the CO-exposed deoxy-HbA. UVRR spectra were acquired at 6 °C. UVRR spectra were normalized at the perchlorate standard peak (934 cm¹) prior to calculation of a UVRR difference spectrum.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

Fig. 1a shows the 229-nm excited UVRR spectrum of human adult deoxy-HbA under solution conditions (7). Both tyrosine and tryptophan bands predominate. The former are located at 1177 cm¹ (Y9a), 1208 cm¹ (Y7a), and 1617 cm¹ (Y8a), whereas the latter are located at 1010 cm¹ (W16), 1236 cm¹ (W10), 1340 cm¹ and 1360 cm¹ (W7), 1459 cm¹ (W5), 1493 cm¹ (W4), and 1559 cm¹ (W3). Changes in several of these UVRR bands occur upon ligand binding. Of considerable significance (see below) is the decrease in intensity of the W3 shoulder at 1549 cm¹ and the frequency shift of the Y8a band to 1616 cm¹. Fig. 1b illustrates these changes in a UVRR difference spectrum, deoxy-HbA minus CO-HbA, under solution conditions. Several features in the difference spectrum have been characterized and are associated with the T-to-R state conformational changes involving tryptophan and tyrosine residues (7). The UVRR difference spectrum for sol-gel encapsulated HbA in different ligation states at 6 °C is given in Fig. 1d. Here, the minuend is the UVRR spectrum for deoxy-HbA, whereas the subtrahend is the spectrum for the same deoxy-HbA sol-gel sample 1 h after the introduction of CO. It should be noted that the reduced signal-to-noise ratio of Fig. 1d is because of the shortened signal acquisition time of 5 min. The short acquisition time was used to preclude, or at least to minimize, the impact of an evolving spectrum for the ligand bound species. For stable equilibrium species, signal acquisition times (Fig. 1b, e.g.) are on the order of 30 min.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   UV resonance Raman (229 nm excitation) spectrum (a) and difference spectra (b-d). a, deoxy-HbA in solution; b, deoxy-HbA minus CO-HbA, samples under solution conditions: 0.125 mM deoxy-HbA in 50 mM Hepes, 0.2 M perchlorate buffer at pH 7.5 at 6 °C; 20 min spectral acquisition time. Sol-gel preparations: deoxy-HbA minus deoxy-HbA 5 h, 20 min after CO ligation; sample at 25 °C after 1 h (c); deoxy-HbA minus deoxy-HbA 1 h after CO ligation at 6 °C (d). Sol-gels bathed in 50 mM Hepes, 0.3 M perchlorate at pH 7.5; samples prepared with a 6-fold excess of IHP.

The UVRR difference spectrum for the sol-gel encapsulated hemoglobin (Fig. 1d) reveals that after 1 h of CO exposure, the intensity of tryptophan bands W16 at 1006 cm¹ and W3 at 1558 cm¹ has substantially decreased. The yet unassigned 1512 cm¹ band has also decreased in intensity. Tyrosine peaks at 1177 and 1620 cm¹ are on the same order of magnitude as noise. The broad band at 1040 cm¹ is a base-line subtraction artifact. Because the intensity differences in the UVRR difference spectrum in Fig. 1d are evident as a limited subset of well separated Raman bands, they cannot be ascribed to artifact. Given the absence of any strong differences in tyrosine bands throughout the full spectrum, the difference feature at 1608 cm¹ is tentatively assigned to a W1 mode (21).

The UVRR difference spectrum in Fig. 1c illustrates subsequent conformational changes in the CO-saturated, encapsulated HbA (subtrahend) vis-à-vis deoxy-HbA (minuend). After the first hour at 6 °C, the CO-exposed sample was then warmed to 25 °C for 4 h to facilitate conformational mobility (1) and then re-cooled to the initial 6 °C for spectroscopic examination. After this warming cycle, the UVRR difference spectrum exhibited several new features. A shoulder at 1549 cm¹ has been added to the 1559 cm¹ peak. The Y8a derivative-shaped difference peak from 1597 to 1627 cm¹, prominent in the solution difference spectrum (Fig. 1b), is still not discernible; however, a decrease in intensity at 1620 cm¹ is apparent. The same observation holds for the Y9a difference peak at 1181 cm¹.

Absorption spectra corresponding to the three states of the encapsulated HbA (discussed above) are very similar to spectra obtained from corresponding solution phase species, indicating that deoxy-HbA and, later, CO-HbA predominate in the sol-gel sample (supplemental data available upon request). The deoxy-HbA absorption spectrum was taken subsequent to the sol-gel wet-aging (20 h), indicating that deoxy-HbA is stable in sol-gel. The UVRR spectra of deoxy-HbA and CO-HbA in sol-gel alone are indicators of the globin conformational state as the W3 and Y8a bands are sensitive to Trp and Tyr conformation, which report on ligation state-associated globin tertiary and quaternary changes (7, 21).

A comparison between the solution phase difference spectrum (Fig. 1b) and those for encapsulated HbA (Fig. 1, c and d) reveals substantial encapsulation-associated differences. These differences cannot be ascribed to a mixture of the equilibrium T and R state species that give rise to the Fig. 1b solution phase differences.

In solution, the binding of CO to deoxy-HbA initiates an undetermined sequence of rapid tertiary structure changes that ultimately result in a quaternary switch from the low affinity T structure to the high affinity R structure (22). The deoxy-HbA minus CO-HbA solution phase UVRR difference spectrum (Fig. 1b) is primarily reflective of changes in the R-T sensitive ₁₂ dimer interface arising from differences between the deoxy T and liganded R conformations (7, 23-29). The W3 difference feature at 1549 cm¹ (Fig. 1b) on the low frequency side of the central peak at 1559 cm¹ (Fig. 1a) originates from a T-R conformational difference in Trp37, the pivotal residue in the hinge region of the ₁₂ interface (7, 23, 26, 28). Whereas Trp37 is responsible for the low frequency shoulder of the W3 band, Trp14 and Trp15, located on the A helices, are the source of the central W3 peak at 1559 cm¹ (7, 21). As the solution phase W3 difference spectrum (Fig. 1b) shows, the two A helix tryptophans and their environment do not change upon switching from deoxy T to liganded R conformations (7, 21). Changes in the intensity of the central peak of W3 for photodissociated and partially liganded intermediates (21, 24), fluoromet-hemoglobins (24), 6 mutants (20, 30, 31), and chemically modified hemoglobins (32) have been ascribed to alterations in the packing of the A and E helices. This conclusion follows from x-ray crystallographic results which indicates that both 14 and 15 are hydrogen bonded to residues on the E helix (33).

The derivative-shaped feature at 1597-1627 cm¹, associated with Y8a (Fig. 1b), has been shown to originate from Tyr42 in the switch region of the ₁₂ interface (27, 29, 34). The hydrogen bond between Tyr42 and Asp99 is broken upon switching from the T to R quaternary structure (35). Intensity changes in Y8a without frequency shifts have been linked to variations in the packing of the penultimate tyrosines, 140 and 145, which alter their spectra-determining hydrogen bonding to the carbonyl of the respective FG5 valines (24, 27, 29, 35) as well as interactions with surrounding residues and solvent molecules.

Fig. 1, c and d, clearly reveals that the binding of CO to deoxy HbA in the sol-gel initiates a discernible sequence of changes. The first changes (Fig. 1d) involve primarily, if not entirely, tryptophan bands. In particular, the central portion of W3 at 1559 cm¹ shows a dramatic decrease in intensity, indicative of a weakening of the hydrogen bonds between the A helix tryptophans and their respective partner on the E helix. Noticeably absent is a change in the low frequency shoulder of W3 arising from Trp37 (Fig. 1b).

The absence in the 1 h sol-gel UVRR difference spectrum (Fig. 1d) of strong derivative-shaped Y8a and Y9a bands seen in the solution-phase difference spectrum (Fig. 1b) indicates that the T state hydrogen bond between Tyr42 and Asp99 is still largely intact. Both the lack of intensity change and frequency shift in the Y8a and Y9a bands for the sol-gel HbA at early time (Fig. 1d) implies that the packing of both penultimate tyrosines is unperturbed (29).

Fig. 1, b-d, implies a progression of conformational changes for the deoxy-Tliganded R transition that starts with a change in the distal heme pocket, as signaled by the Trp W3 (1559 cm¹) and W16 (1010 cm¹) difference bands (Fig. 1d). It can be seen (Fig. 1c) that at longer times (and exposure to higher temperatures) the encapsulated, CO-exposed deoxy-HbA sample has undergone additional changes. The appearance of the 1549 cm¹ difference band (Fig. 1c), signals a shift of the hinge region Trp37 in the ₁₂ interface toward its R state configuration. The decrease in the intensity of the Y8a and the Y9a peaks (Fig. 1c) are ascribed to a loosening (increased disorder) of one or both penultimate tyrosines.

Crystallographic studies (36) on a series of 37 mutants provide a plausible structural and mechanistic explanation for the spectral changes shown in Fig. 1c that is based on the coupled behavior of Trp37 and Tyr140. The packing of Tyr140 is likely to be disrupted by the ligand binding-induced shift in the F helix because of movement of the iron into the plane of the heme porphyrin. Arnone and co-workers (36) describe a lever-like amplification of the iron motion via the vertical displacement of the proximal histidine F8 (His87) and the adjacent F9 Ala88 that directly contacts the phenol ring of Tyr140. The movement of this lever upon ligand binding should result in increased disordering for the tyrosine side chain. The tight packing of the tyrosine side chain in the deoxy-T state maintains the T state 37 hinge configuration. An increase in the mobility of the tyrosine side chain allows the environment about 37 to slip toward its R state conformation, thus linking the tyrosine- (Y8a, Y9a) and Trp37 (W3)-associated spectral features seen in the "5 h" UVRR difference spectrum (Fig. 1c). Visible Raman spectroscopic (37) and x-ray crystallographic (36) studies on 37 mutants imply both that the proximal heme environment is tightly coupled to these degrees of freedom and consequently, that the above UVRR changes should be accompanied by a decrease in proximal strain at the  hemes. The progressive increase in the iron-proximal His visible Raman frequency in a series of T state Trp37 mutants implies a progressive decrease in proximal strain and is correlated with both an increased disorder in 140 and a corresponding enhancement in ligand binding rates (37, 38).

In conclusion, the encapsulation of deoxy-HbA with IHP in porous sol-gels has enabled us to study a homogeneous population of ligand-bound deoxy-HbA. This accomplishment is a direct consequence of the ability of the sol-gel to slow the rate of conformational change relative to that of ligand diffusion within the sol-gel. The UVRR difference spectra obtained from the binding of CO to encapsulated deoxy-HbA reveal an initial E helix motion upon ligand binding tentatively attributed to the  subunits. The subsequent simultaneous appearance of Tyr (Y8a and Y9a) and W3 (1549 cm¹) UVRR difference bands is ascribed to a ligand binding-induced increase (via the iron-F8 proximal histidine-F9Ala lever) in the disorder of the penultimate Tyr140 and the concomitant shifting of the "hinge" region Trp37. Studies designed to probe the next phases of relaxation are in progress. More generally, this report demonstrates that sol-gel encapsulation of proteins can be used to study sequences of conformational events triggered by substrate binding that are not accessible via conventional rapid mixing experiments.

	FOOTNOTES

^* This work was supported by the National Institutes of Health Grants PO1 HL51084 and RO1 HL58247 and by the W. M. Keck Foundation.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The on-line version of this article (available at http://www.jbc.org) contains Figure S1.

Professor in the Department of Physiology and Biophysics. To whom correspondence should be addressed: Albert Einstein College of Medicine, 1300 Morris Park Ave., Bronx, NY 10461. Tel.: 718-430-3591; Fax: 718-430-8819; E-mail: jfriedma@aecom.yu.edu.

	ABBREVIATIONS

The abbreviations used are: UVRR, ultraviolet resonance Raman spectroscopy; HbA, adult human hemoglobin A; TMOS, tetramethylorthosilicate; CW, continuous wave; IHP, inositol hexaphosphate.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES