Effect of Hydrostatic Pressure on Spectra of Heme Compounds*

Increase in hydrostatic pressure shifts the absorption bands of oxy-, carboxy-, and deoxyhemoglobin and myoglobin toward the red by 0.4 to 0.7 nm corresponding to a change in extinction coefficient of from 4 to 8% at the peak of the difference spectrum. The pressure difference spectrum for oxyhemoglobin closely resembles the difference spectrum described by Adams and Schuster ((1974) Biochem. Biophys. Res. Commun. 58, 528-533) following addition of inositol hexaphosphate to oxyhemoglobin. A similar shift was observed for derivatives of dimethyl-deuterohemedisulfonate in both Fe2+ and Fe3+ forms indicating that the protein is not required for this effect, in contrast to earlier reports of T. L. Fabry and J. W. Hunt ((1968) Arch. Biochem. Biophys. R3, 428-429) and Q.H. Gibson and F.G. Carey ((1975) Biochem. Biophys. Res. Commun. 67, 747-571) who were unable to observe changes in aqueous solutions of protoheme derivatives.


Effect of Hydrostatic
Pressure on Spectra of Heme Compounds* ( Increase in hydrostatic pressure shifts the absorption bands of oxy-, carboxy-, and deoxyhemoglobin and myoglobin toward the red by 0.4 to 0.7 nm corresponding to a change in extinction coefficient of from 4 to 8% at the peak of the difference spectrum. The pressure difference spectrum for oxyhemoglobin closely resembles the difference spectrum described by Adams and Schuster ((1974) Biochem. Biophys. Res. Commun. 58, 528-533) following addition of inositol hexaphosphate to oxyhemoglobin. A similar shift was observed for derivatives of dimethyldeuterohemedisulfonate in both Fe*+ and Fe"+ forms indicating that the protein is not required for this effect, in contrast to earlier reports of T. L. Fabry and J. W. Hunt ((1968) Arch. Biochem. Biophys. R3, 428-429) and Q. H. Gibson and F. G. Carey ((1975) Biochem. Biophys. Res. Commun. 67, 747-571) who were unable to observe changes in aqueous solutions of protoheme derivatives.
Many studies of hemoproteins have made use of their characteristic absorption spectra which are both intense and widely different for different hemoproteins and for various derivatives of the same hemoprotein. In most cases it has been assumed that the spectrum of a given compound is independent of solution conditions, such as pH, pressure, temperature, and buffer composition, and this is certainly a good first approximation.
More detailed studies have shown, however, that some variation in the spectrum of individual derivatives may occur, and the two best known examples are deoxy-and oxyhemoglobins.
Deoxyhemoglobin occurs in two forms, one of which has a stronger and sharper Soret band than the other. The stronger band, seen in normal mammalian deoxyhemoglobin, is associated with a low ligand affinity and may be interpreted as showing that the hemoglobin is in the T state of the Monod-Wyman-Changeux model (1). The form with the weaker Soret band is associated with high affinity and the R state of the model. This correlation is, so far, without exception and has been extended to many examples as discussed in *This research was generously supported by National Science Foundation Grants BMS 74.083233 and GB 38272 and United States Public Health Service Grant GM 14276. Contribution 3903 and 3904 of the Woods Hole Oceanographic Institution. the recent review of Baldwin (2), although this is not concerned with spectra as such.
The position for oxyhemoglobin is much less clear, but Adams and Schuster (3) have recently described a characteristic difference spectrum on adding inositol hexaphosphate to hemoglobin A, which they suggested might be due to an R to T transition of the liganded form. Since then, similar spectral changes have been observed as a result of change in pH with carp carboxyhemoglobin, and on changing the temperature of oxyhemoglobin A (4). One component of trout hemoglobin also responds similarly to pH change (5). While there is good reason to think that carp carboxyhemoglobin may change from the R to T conformation on lowering the pH (6,7), there is less reason to suppose that changing the temperature alters the conformation of a series of hemoglobin A derivatives, and the specificity of the Adams and Schuster spectrum is thus open to some question. Some years ago Fabry and Hunt (8) gave a preliminary account of studies of the effect of pressure on the absorption spectrum of several hemoproteins and heme derivatives, finding effects on the Soret but not on the visible spectrum for hemoglobin, or on protohematin. More recently Zipp and Kauzmann (9) have made detailed studies of the effect of pressure on the denaturation of metmyoglobin, and have reported the associated spectral changes, but in neither case was the work designed to explore a possible relation between the spectrum and the R-T transition.
This paper describes the effect of increased hydrostatic pressure on the spectrum of several hemoglobin derivatives. It appears that pressure will produce changes analogous to those observed by Adams and Schuster, and that the protein is not necessary for this to occur.

EXPERIMENTAL PROCEDURES
Materials -Human hemoglobin was from hemolyzed human blood obtained bv caDillarv Duncture. Menhaden (Breuoortia tvrannus) blood was obta&ed frb& the caudal vein of several fish. The-red cells were washed twice with saline and stored in liquid nitrogen. This frozen material was thawed in 20 volumes of distilled wat& and the gelatinous precipitate centrifuged off before use.
Dimethyldeuterohemedisulfonate was the gift of Dr. R. L. J. Lyster, National Institute for Research in Dairying, Reading, England. Compressed gasses were from Matheson Co. Other chemicals were reagent grade.
Pressure Chamber-The pressure chamber was bored from a 7.5 cm high, 5.6 cm long, 4 cm wide block of 303 high tensile strength stainless steel, and had a l-cm light path and a 2.5-ml volume. The windows were 1 cm thick fused silica with a thickness to unsupported diameter ratio of 1:3. They were held in stainless steel plugs which screwed into the body of the chamber.
Pressure was applied with a hand-operated hydraulic pump (Blackhawk model P228, 40,000 p.s.i.) through i/4 inch outer diameter, i/i&h inch wall stainless tubing connected with swagelock fittings.
A vertically mounted steel cylinder, 1.3 cm inner diameter, 15 cm long, was placed in line between the pump and the pressure chamber to serve as an oil-water separator, the system distal to this being filled with distilled water. A free piston in the pressure chamber separated the hydraulic system from the experimental solution and a separate filling port provided ready access for changing the solution.
Spectrophotometry-With the Cary USC, groups of spectra which were to be compared with one another were recorded without releasing the drive gear train to preserve wavelength calibration. Absolute spectra were recorded in every case. These were measured at intervals of 2 nm; difference spectra were generated, and corrections for path length and concentration change were made after transferring the readings representing the spectra to an 8/I computer (Digital Equipment Corp.). In addition to the spectra of the hemoglobin derivatives themselves, base-lines for solvent at 1,100, and 1000 atm were recorded.
These reflect changes in quantities such as stressinduced birefringence in the windows rather than changes in the solvent itself. Analysis of the data suggested that absorbance values had an error of about ?0.005.
Direct digital recording of spectra was performed using a laboratory-constructed instrument which has been described in detail elsewhere (Knowles and Gibson (10)). This is a simple split-beam spectrophotometer based on a Bausch and Lomb 500 mm f.  Fig. 1 shows the absolute spectrum recorded for oxyhemoglobin over the range 400 to 460 nm, and Fig. 2 represents the same data expanded for the range 410 to 420 run only. It is clear from Fig. 2 (4). In using data from plots such as that of Fig. 3   3. A comparison of the difference spectrum produced by 1000 atm hydrostatic pressure (0) and an arbitrary difference spectrum generated by a numerical shifting of the wavelength by 0.46 nm (A). The experimental data are taken from Fig. 1. the shift in the position of a peak in the absolute spectrum, which is closely true for small shifts such as those considered here. The shift in the wavelength of maximum absorbance on applying increased hydrostatic pressure was calculated by comparing the effect of a 1-nm shift with that actually observed. Marked asymmetry of the difference spectrum is an indication of a change in intensity of the heme absorbance, and in the few cases where this occurred a note has been added to the data collected in Table I.  Table I also includes the peak to peak amplitude of the pressure difference spectrum expressed as a percentage of peak amplitude. This figure obviously depends upon the shape of the absorption bands and is included only to give a concrete idea of the meaning of the figures for the band shift included in Table I. Effect of Pressure on Spectrum of Several Heme Derivatives -The compounds listed in Table I include several derivatives of human and menhaden hemoglobin and of dimethyldeuterohemedisulfonate.
The data for these compounds were collected in a form which allowed their transfer to the computer and analysis as described in the previous paragraph. In addition, CO myoglobin, deoxymyoglobin, and n-butylisocyanide hemoglobin were examined using the Cary 118C spectrophotometer; the CO compound showed a clear shift similar to that for human hemoglobin, but no effect on deoxymyoglobin was seen in the visible region. The effect of pressure on the spectrum of n-butylisocyanide hemoglobin and myoglobin is also to produce a shift to the red of about 0.5 nm. A number of experiments were also performed using several protoheme derivatives.
These were technically much less satisfactory than those with hemoproteins, with a tendency for pressure to produce slowly reversible shifts and changes in intensity of the spectrum. Shack and Clark (12) long ago established that several hematin compounds tend to be polydisperse in aqueous Shift in absorption band due to pressure Absorption spectra with about 5 PM heme in Soret, 50 pM in visible region, l-cm path, 20", pressure change 1000 atm, 500 pM inositol hexaphosphate, 3 mM n-butylisocyanide, 250 pM O,, 900 pM CO, and 1 mM CN. This compound is freely soluble in water at neutral pH, and has sharper and better defined bands than protoheme. With it, clear shifts were easily demonstrated on increasing the hydrostatic pressure of about the same amplitude as shown by the hemoproteins. A single experiment with reduced cytochrome c showed no effect of pressure on the o( band. DISCUSSION The results described here confirm the report of Fabry and Hunt (8) that increased hydrostatic pressure shifts the Soret absorption bands of several heme derivatives toward the red.
In addition, we have been able to observe similar shifts in the position of the visible absorption bands except for deoxyhemoglobin. In our first experiments (13), like Fabry and Hunt, we were unable to find a clear effect of pressure on protohematin, and so thought that the protein was necessary to obtain the effect. In experiments with the freely water-soluble dimethyldeuterohemedisulfonate, however, shifts of about the same extent were observed with increased pressure for a series of derivatives both ferrous and ferric, showing clearly that the effect is upon the heme, not upon the protein. In fact, the only compounds which failed to show an appreciable effect were deoxyhemoglobin and reduced cytochrome c. It is tempting to speculate that the effects are really due to changes in the character of water which, on increase in pressure, loses partially ordered elements to less bulky random arrangements, and so increases in effective polarity as a solvent. The effect of pressure would then be a solvent effect exerted on the r -z-* transition which is associated with the principal absorption bands of heme compounds. However this may be, the experiments described here seem to establish that the Adams and Schuster difference spectrum or an analogue of it may be obtained from almost all the heme compounds tested by the application of hydrostatic pressure. The spectrum is closely imitated in every case by a synthetic spectrum formed by numerical differentiation of the spectrum of the appropriate compound at 1 atm pressure. The main exception was deoxyhemoglobin where no shift could be seen in the visible region, although a normal shift was observed in the Soret. Numerical experiment suggests that our methods should have detected a shift in the visible band, and we have no explanation for our failure.
The range of compounds showing the effect is such as to make it clear that the R-T transition of the Monod-Wyman-Changeux model (1) is in no way involved, and with the menhaden hemoglobin it was possible to observe quite similar effects of pressure on both R and T forms of deoxy and liganded hemoglobins. The population of the R and T forms was judged by functional criteria in ligand binding.l Although our observations do not explain the origin of the changes seen by Adams and Schuster (31, they do show that these changes cannot be regarded by themselves as diagnostic of the occurrence of an R-T transition in solution.
The effects on the spectrum are not large enough to make it easy to establish an accurate relation between pressure and spectral shift, and it is not possible to say if we are observing a large effect on a few molecules or a small effect on most of the molecules in solution. So far as our observations permit an estimate, it seems that the effect increases in proportion with 'F. G. Carey and Q. H. Gibson, unpublished observations. pressure. Since it is possible to obtain effects with soluble heme derivatives it might be of interest to an investigator with apparatus able to operate at higher pressures to see to what point the effects will continue to increase, since with these compounds no limitation due to protein denaturation will arise.
There is, finally, the matter of the practical significance of these shifts in absorption spectrum. Although the shifts are not very large, the error introduced in the determination of a ligand binding curve at high pressure could be very considerable, and would vary significantly with the observing wavelength used. The effect under ordinary conditions is not appreciable.