THEJOURNAL CHEMISTRY

OF BIOLCGICAL Vol. 269, No.49, Issue of December 31047-31050,1994

0 1994 by The American Society for Biochemistry and MolecularBiology, Inc. Printed in U.S.A.

Low Frequency Vibrational Modes of Oxygenated Myoglobin,

Hemoglobins, and Modified Derivatives*
(Received forpublication, August 4, 1994, and in revised form, September 19, 1994)

Shanthini JeyarajahS, LeonardM. ProniewiczSOn, Helen BronderS, and JamesR. KincaidSll

From the Wepartment of Chemistry, Marquette University, Milwaukee, Wisconsin 53233 and the $Regional Laboratory of
Physicochemical Analysesand StructuralResearch, Jagiellonian University, ul. Ingardena 3, 30-060 Krakow, Poland

The low frequency resonance Raman spectra of the Thus, intensive RR studies have been undertaken to identify
dioxygen adducts of myoglobin, hemoglobin, its isolated the internal modes associated with these fragments.
subunits, mesoheme-substituted hemoglobin, and sev- In the case of the CO adduct of hemoglobin, dFe-CO), 6(Fe-
eral deuteriated heme derivatives are reported. The ob- C-0) and the internal 4CO) modes have been assigned by
served oxygen isotopic shifts are used to assign the careful iron- isotopic labeling studies(8).On the other hand,for the
oxygenstretching (-570 cm”) andtheheretofore most physiologically important derivative, the dioxygen ad-
unobserved 6(Fe-0-0)bending (-420 cm”) modes. Al- duct, only the 4Fe-0,) has thus far been identified (121, al-
though the6(Fe-0-0)is not enhancedin the caseof oxy- though we have previously presented evidence for the enhance-
myoglobin, it is observed for all the hemoglobin deriva- ment of the bendingmode, S(Fe-O-O),at 491cm” in thecase of
tives, its exactfrequencybeingrelativelyinvariable the 0, adduct of ferrous lactoperoxidase (13). Brunner first
among the derivatives. The lack of sensitivity to K O / reported a band at 568 cm” in theRR spectrum of HbO,, which
D,O buffer exchangeis consistent withour previous in-
exhibited the expected isotopic shift of 4Fe-0,) upon “0, sub-
terpretation of H,O/D,O-induced shifts of 140-0) in the
resonance Raman spectra of dioxygen adducts of cobalt- stitution (12). Althoughthe assignmentof this mode to 4Fe-0,)
substituted heme proteins; namely, that those shifts arewas subsequently questioned by Yu and co-workers (ll),who
associated with alterations in vibrational coupling of preferred assigning it to the bendingmode 6(Fe-0-0) based on
u ( 0 - 0 ) with internal modes of proximal histidyl imidaz-isotopic shift patterns, Nakamoto and co-workers (141, using
ole rather than to steric or electronic effects of W D ex- similar datafor the 0, adduct of iron phthalocyanine,provided
change at the active site. No evidence is obtained for convincing arguments t o support the original assignment.
enhancement of the u(Fe-N) stretching frequency of the Obviously, the iron-imidazole linkage to proximal histidine
linkage between the heme iron and the imidazole group has attracted much attention in the study of the cooperative
of the proximalhistidine. mechanism because of its expected sensitivity to changes in
quaternary structure and solution conditions (15, 30). In the
deoxy forms of hemoglobin and myoglobin the Fe-NJimidazole)
stretching vibration has been assigned definitively by the shifts
Resonance Raman (RR)’ spectroscopy is firmly established
observed upon 15N/14Nand 57Fe154Fe substitution aswell as by
as an effective probe of structural changes associated withco-
operative ligand binding tohemoglobin (1,2). The vibrational 2H,0iH,0 buffer exchange, which serves to probe the deuteri-
modes associated with the heme macrocycle, its peripheral sub- ated proximal histidine (16, 17). Although the corresponding
stituents, and the axial ligands have been detected and as- mode for a ligated heme protein has not been identified unam-
signed. The strongly enhanced in-plane modes of the heme biguously thus far, tentative assignments have
been proposed
macrocycle have been analyzed extensively, and reliable “mark- for weak features observed a t 272 cm” and 309 cm” in the
er bands”of heme core size, spin state, and oxidation state have spectra of oxymyoglobin (18)and cyanomet hemoglobin CTT I11
been identified (3-5). Selective isotopic labeling and hemeside (181, respectively.
chain modification studies haveprovided some insight into the In this work we report the results of RR studies of oxygen-
contributions from peripheral substituents (6, 7, 19). ated hemoglobin, wherein we have made an intensiveeffort to
The vibrational modes associated with the iron-axial ligand locate and identifymodes associated with the Fe-0, andFe-N,
fragments, on both the proximal and distal sides, obviously linkages. Inasmuch as these are generally weak features over-
have great potentialfor elucidating detailed bonding informa- lapped with heme macrocycle modes, it is obviously difficult to
tion relevant to cooperative ligand binding. On the distalside, observe distinct spectral evidence for these particular modes.
heme binds a variety of exogenous ligands including NO, CO, Nevertheless,careful comparison of spectraobtainedusing
CN-, and 0,, and extensive efforts have been made t o under- multiple isotopic labeling as well as a chemically modified
stand thebonding in theseiron-ligand fragments (8-11,28,29). heme (mesoheme) providesconvincing evidence for assignment
of the heretofore unobserved S(Fe-0-0) mode as well as the
absence of enhancement of the u(Fe-N,).
* This work was supported in part by National Institutes of Health
Grant DK13153. The costs of publication of this article were defrayed in EXPERIMENTAL PROCEDURES
part by the payment of page charges. This article must therefore be
hereby marked “advertisement” in accordance with 18 U.S.C. Section Materials and Methods-Hemoglobin A was isolated from washed
1734 solely to indicate this fact. red blood cells obtained from a local blood bank according to literature
llSupported by Grant 2P 303 060 05 from the Polish Committee for procedures (20). The oxyhemoglobin was prepared by saturating the
Scientific Research. deoxyhemoglobin with 0, at atmospheric pressure. The isolated sub-
11 To whom correspondence should be addressed: Dept. of Chemistry, units were prepared as described previously (21).
MarquetteUniversity,535 N. 14th St., Milwaukee, WI 53233.Tel.: The mesohemoglobin was prepared by titrating apoglobin solution
414-288-7065;Fax: 414-288-3300. withmesohemin.The mesohemin was purchased from Midcentury
The abbreviations used are: RR, resonance R a m a n ; bis-tris, 2-bis(2- (Posen, IL). Apoglobin was prepared from carbon monoxyhemoglobin
hydroxyethyl)aminol-2-(hydroxymethyl)-pmpane-l,3-dio~ Mb, myoglobin. via the acid acetone method. The resulting white precipitate was dis-

31047
31048 Resonance Raman Spectra of Oxyhemoglobin
solved in a minimum amountof distilled water (200 mg of globin in 10
ml), and the residual acetone solution was removed by dialyzing the
globin solution in waterfor 0.5 h. It was then dialyzed against bis-tris,
pH 6.0, for 2 h. Theapoglobin solution was then titrated with mesohe-
min a s reported previously (21).
The resulting solution was allowed to stand overnight a t 0 "C. The
excess hemin was readily removed by running on a Sephadex G-25
(Pharmacia Biotech Inc.) column (2.5 x 40 cm) equilibrated with 20m~
phosphate buffer, pH 6.0. Excess KCN was then added to obtain the
metcyano derivative,and thesolution wasallowed to stand for 0.5 h in
an ice bath. The excessKCN was removed by passing through another
Sephadex G-25 column of the same dimensions equilibrated withmM 20
phosphate buffer, pH 7.0.
The metcyanomesohemoglobin was passed through aBio-Gel P-6DG
(Bio-Rad) column equilibrated with 20 mM phosphate buffer, pH 6.4,
and loaded onto a CM-cellulose column equilibrated with the same
buffer. Thepuremethemoglobin, which bindstothecolumn,was
washed with the same buffer and then eluted with 50 m~ phosphate
buffer, pH 7.4. The eluent was then concentrated to the desired volume
by using Centricon CM-30 (Amicon) microconcentrators. The sample
was then reduced by the additionof sodium dithionite undera nitrogen
atmosphere and immediately passed through Bio-Gela P-GDG column
equilibrated with 0.05 M phosphate buffer, pH 7.5. The mesohemoglobin
was collected as the oxygenated derivative, andthe purity waschecked
usingW-visible spectroscopy. The mesodeuteriated hemoglobin
(HbO,d,) was prepared in the same manner described above using me-
sodeuteriated (a,p, y, 6) protohemin.
The substitution of I6O2by '*O, was carried out on a vacuum line
where the protein sample was first degassedby the vacuum and then
saturated with N,. This procedure was repeated until all the 1602 was
removed from the sample. The evacuated tube was exposed then to '*02.
The 2H,0/H,0 buffer exchange was carried out in the microconcen-
trator. About 1ml of HbO, was concentrated to 0.2 ml and diluted with
'H,O buffer to the original volume. This procedure was repeated three
times to ensure complete exchange.
Horse heart myoglobin (Sigma) was purified by ion-exchange chro- FIG.1. RR spectra of the oxygen adducts of hemoglobin
matography on a column of CM-52 (Whatman). Samples in 10 mM derivatives.
phosphate buffer, pH 6.0, were loaded onto a column equilibrated with
the samebuffer. After the column was washed with approximately200 and the corresponding difference spectra ('60,-'80,) are given
ml of the pH6.0 buffer, the retainedmyoglobin was eluted with 25 m~
in Fig. 2. The absolute and difference spectra for oxymyoglobin
phosphate buffer, pH 8.5. The oxygenated form was generated by re-
duction with sodium dithionite followed by passage over a Sephadex are given in Fig. 3. In addition to the relatively strong differ-
G-25 column as was described above for the hemoglobin derivatives. ence feature observed near 570 cm" in each figure, a second
Raman Measurements-Resonance Raman spectra presented in Fig. weaker difference feature is observed in the -410-430 cm"
1were obtained using a Spex 1269 spectrometer equipped awith Prince- region for all proteins except oxymyoglobin. The strong feature
ton InstrumentsICCD-576 W-enhanced detector anda 413.1-nm notch having positive and negative components at -570 and 545 cm"
filter (Kaiser Optical Systems, Inc., Ann Arbor, MI). Those shown in is readily assigned to the 4Fe-0,) mode, in agreement with
Figs. 3 and 4 were acquired with a Spex 1403 spectrometer equipped
with a Hamamatsu R-928 photomultiplier tube and a Spex DMlBcon- earlier studies(12, 18).The weaker set of features occurring in
troller. All samples were excited with the 413-nm line availablefrom a the 410-430 cm-' region are most reasonably assigned to the
Coherent Innova 100 krypton ion laser except mesohemoglobin, which heretofore unobserved XFe-0-0) mode. It is important to men-
was excited using the 406.7-nm line from the same laser. A spinning tion that these weak features are not readily apparent in the
NMR tube was used to record all the spectra at room temperature. A absolute spectra because of overlap with heme modes near 430
highly concentrated sampleof native hemoglobin (5 mM) has been used and 410 cm". Thus, the only manifestation of these in the
in our study inasmuch as the iron-imidazole stretching gives such
weaker RR scattering in the oxy form. Typical concentrations ranged
absolute spectra are the subtle changes in intensities in this
from 0.5 to 1 mM for the modified hemoglobins. All spectra were cali- region.
brated by scanning over the Rayleigh linewith low collection efflciency. Although a similar mode has not been identified previously
The samples were fully oxygenated as monitored by the absence of 1358 for the oxygen transport heme proteins, a second oxygen iso-
cm" band (va of deoxyhemoglobin) and by the presence of 1378 cm" tope-sensitive feature has been observed for the dioxygen ad-
band (v4 of oxyhemoglobin). Complete substitution of was obtained duct of ferrous lactoperoxidase (13)as well as for several model
in allof our experiments as is evident fromthe shiftof the band at570
cm" to 545cm" (Le. there wasno residual intensityat 570 cm"). All the compounds (14, 22). We note that the frequency observed for
spectra were taken under the same experimental conditions, and laser HbO, is approximately intermediate between the 345 cm" ob-
power was maintained a t about 20 mW using a cylindrical lens to avoid served for 6(Fe-0-0) of the five-coordinate FeTPP.0 (22) and
thermal decomposition of the samples. The slit width was maintainedthat at observed at 491 cm" in the case of lactoperoxidase com-
2-5cm" correspondingto a mechanicalslitwidth of 150-300 pm. pound I11 (13), which is known to havea stronger iron-histidine
Reported wave numbers are accurate tof 1 cm". bond than does hemoglobin.
RESULTS AND DISCUSSION Structural and EnvironmentalZssues
Vibrational Modes Associated with the Fe-0, Fragment Distal Side Hydrogen Bonding Effects-One of the most often
The low frequency RR spectra of the dioxygen (I6O,) adducts discussed issues involving dioxygen binding to heme proteinsis
of native hemoglobin, its isolated CY and p subunits and meso- the extent to which a hydrogen bonding interaction between
heme-substituted hemoglobin are given in Fig. 1. The most the bound dioxygen and the imidazole NH group of the distal
effective approach t o identify the weak features associated with histidine affects the Fe-0,linkage, and, indeed, the vibrational
the Fe-0, fragment is toemploy isotopically labeled dioxygen, modes associated with the Fe-0, fragmentmay be expected to
 ResonanceSpectra
Raman
m
u3
0
10
2
d
c
4"
u3
D d u l
FIG.2. Difference spectral ('o0,'80,).A, H b ; B , 01 subunits; C , p
subunits; D , mesoheme-substituted hemoglobin.
be sensitive to this interaction. In fact, slight frequency shifts of
the 40-0) mode in the RR spectra of 0, adducts of cobalt-
substituted hemoglobin and myoglobinupon H,O/D,Oex-
change had earlier been interpreted in termsof slight changes
in the strength of this hydrogen bond upon substitution of
deuterium (24). These shifts were subsequently confirmed but
were instead attributed to alterations in vibrational coupling
interactions between the do-0) and internal modes of the
proximal histidyl imidazole (25). Thus, the H,O/D,O shifts
could be quantitatively accounted for by assuming absolutely
no effect on the inherent frequency of the 40-0)(26).
We have acquired the low frequency RR spectra of the 0,
adducts of hemoglobin in both H,O and D,O buffers and have
confirmed that there are no resultant difference features ob-
servable. These data areconsistent with our previous interpre-
tation for the cobalt-substituted derivatives; namely, that the
spectral changes observed for 40-0)upon H,O/D,O exchange
are associated with alteration in vibrational coupling rather
than with any steric or electronic effects ascribable to WD
exchange at the active site.
It isalso important to point out that thefrequencies observed
for the 4Fe-0) and S(Fe-0-0)modes of these oxygenated de-
rivatives are consistent with the correspondingmodes observed
for the isoelectronic NO adducts of the cobalt-substituted ana-
logues (23). Thus, in that case the dFe-N) and G(Fe-NO) were
observed at 576 and 367 cm", respectively.
Heme Structure-It is evident from comparisonof the spectra
given in Figs. 1-3 that thegeneral steric and electronic factors
that dictate the frequencies of the features associated with the
Fe-0, fragment are similar for all the proteins studied. On the
other hand, slight differences are noticeable for the other low
frequency modes that are associated with the heme group. In
of Oxyhemoglobin 31049
A - B
200 400 600
Raman shift (em-')
FIG.3. RR spectra of the dioxygen adductsof horse heartmyo-
globin. A, Mbl6O,; B , Mb"0,; C , difference spectrum A - B.
the case of mesohemoglobin (truce D in Fig. 2), it may be ex-
pected that substantial differences in heme mode frequencies
would exist because of the fact that thereplacement of the vinyl
substituents with ethyl groups can obviously lead to slight
changes in mode composition. However, the substantial differ-
ences between the spectra of oxymyoglobin and oxyhemoglobin
as well as themore subtle, butsignificant, differences between
hemoglobin and the isolated subunits deserve attention.
To address this issue we have acquired the low frequency
(200-600 cm") spectra of the 0, adducts of myoglobin and
hemoglobin in which the heme has been replaced with proto-
heme that hasbeen selectivelydeuteriated at thefour methine
positions. The spectra of these deuteriatedderivatives are com-
pared with those of the native species in Fig. 4,inspection of
which reveals that, inboth cases (HbO, and MbO,), the modes
that exhibit the greatest deuterium shifts are those near 550,
430, and 300 cm". It is interesting t o point out that it is pre-
cisely thisset of modeswhichshows greatest variability
among the spectra of the native systems (Figs. 1 and 3). The
region near 300 cm" is particularly noteworthy. Referring to
Fig. 1, in the cases of HbO, and the isolated a! subunits, the
featureinthis regionoccurs near 295 cm".On the other
hand, this mode is observed approximately 15 cm" higher for
MbO, (Fig. 3) and 10 cm" higher for the the isolated p sub-
units (Fig. 1).
Although further experimental work, and eventually a de-
tailed normal mode calculation, would be required to obtain a
clearer understanding of the natureof this mode, its relatively
3 1050 Resonance Raman Spectra of Oxyhemoglobin
(D
n,
d,-MbO,
u3
cu In
200 400
Raman shift (cm-')
FIG.4. RR spectra of the ''Oz adducts. A, hemoglobin A (HbO,); B ,
hemoglobin with methine-deuteriated protoheme (d,-HbO,); C , horse
heart myoglobin (MbO,); D, horseheart myoglobin withmethine-
deuteriated protoheme (d,"bO,).
large sensitivity to methine deuteriation is consistent with its
assignment t o a n out-of-plane heme distortion. We note that a
recent calculation (27) of the out-of-plane modes for a symmet-
ric model compound, NiOEP, predicts a mode (y7)in thisregion
which involves significant motion of the methine carbons and
thus should exhibit sensitivity to deuteriationat this position
(in thecase of NiOEP the observed shift is7 cm"). Inasmuch as
distortions of the equilibrium geometry of hemes and metallo-
porphyrins are most easily accommodated by alterations about
the methine bridges, the variability of this mode may reflect
quitesubtle differences in macrocycle planaritythroughout
this series of otherwise similar structures.
The other notable difference in the RR spectra of the native
systems is the relatively strong enhancementof the modes near
250 and 500 cm-I in the case of MbO,. This behavior has been
noted previously by Champion and co-workers (30) who suggest
that the activation of the -250 cm" mode may be associated
with theangle of orientation of the proximal histidyl imidazole
with respect t o the heme plane,but the data presented here are Yoshikawa, and T.Kitagawa, submitted for publication.
not informative regarding thisissue. However, it isworthwhile
to point out that the general appearance and trendsobserved
here in comparing the heme mode frequencies for HbO, and
MbO, are againconsistent with the data for the isolectronic NO
adducts of the cobalt-substituted analogues (23). Thus, in the
case of NOMb,, a relatively strong 254 cm" feature isobserved
which is not detectedin theRR spectrum of NOHb,,. Similarly,
NOMb,, exhibits a feature near 298 cm" which is found at a
lower frequency for NOHb,,; again as in the case for the 0,
adducts shown here, the single -375 cm" mode (NOMb,,)
d i t s into two comuonents (370 and 383 cm") in the case of
Finaiiy, we note that during the
revision of this manuscript it
came to our attention that Kitagawa and co-workers' have
independentlyobtained evidence for the observation of the
6(Fe-0-0) mode near 430 cm-I and have supported that assign-
ment with a normal mode calculation and additional data ob-
tained using the pure 160180 dioxygen isotopomer.
Acknowledgments-We are grateful to Dr. Songzhou Hu for helpful
discussions and to Professor T. Kitagawa (Institute for Molecular Sci-
ence, Okazaki, Japan)for permission t o cite hiswork prior to publication.
REFERENCES
1. Spiro, T. G. (1983) in Iron Porphyrins (Lever, A. B. P., and Gray, H. B., eds) vol.
2, p. 89, Addison-Wesley, Reading, MA
2. Rousseau, D. L., and Friedman, J. M. (1988) in Biological Applications of
Raman Spectroscopy (Spiro, T. G., ed) Vol. 3, .. pp. 133-215, Wiley
Interscience, New York
3. Choi, S., Lee, J. J., Wei, Y. H., and Spiro, T. G. (1983) J. Am. Chem. SOC.105,
3~12-37n7
...~
4. Spiro, T. G., and Strekas, T. C. (1974) J. Am. Chem. SOC.96, 338-345
5. Yamamoto, T., Palmer, G.,Gill, D., Salmeen, I.T., and Rimai, L. (1973) J. Biol.
Chem. 248,5211
6. Choi, S., Spiro, T. G., Langry, K. C., Smith, K. M., Budd, D. L., and La Mar, G.
N. (1982) J. Am. Chem. SOC.104, 4345-4351
7. Uchida, K., Susai, Y., Hirotani, E., Kimura, T., Yoneya, T., Takeuchi, H., and
Harada, I. (1988) J. Biochem. 103,979-985
8. Yu, N. T., andKerr, E. A. (1988) in Biological Applications of Raman
Spectroscopy (Spiro, T. G., ed) vol. 3, pp. 96-131, Wiley Interscience: New
York
9. Tsubaki, M., Srivastava, R. B., and Yu, N. T. (1982) Biochemistry 21, 1132-
1140
10. Yu, N. T., Benko, B., Kerr, E. A,, and Gersonde,K. (1984) Proc. Natl. Acad.Sci.
U. S. A. 81,5106
11. Benko, B., and Yu, N. T. (1983) Proc. Natl. Acad. Sci. U. S. A . 81, 5106
12. Brunner, H. (1974) Naturwissenchaften 121,61
13. Hu, S.,and Kincaid, J. R. (1991) J . Am. Chem. Soc. 113, 7189-7194
14. Bajdor, K., Oshio, H., and Nakamoto, K. (1984) J. Am. Chem. Soc. 106,7273-
7274
15. Ondrias, M. R., Rousseau, D. L., Shelnutt, J. A,, and Simon, S. R. (1982)
Biochemistry 21, 3428-3437
16. Argade, P. V., Sassaroli, M., Rousseau, D. L., Inubushi, T., Ikeda-Saito, M., and
Lapidot, A. (1984) J . Am. Chem. SOC.106,6593
17. Kitagawa, T.,Nagai, K., and Tsubaki, M. (1979) FEBS Lett. 104,376-378
18. Walters, M. A,, and Spiro, T. G. (1982) Biochemistry 21, 6989-6995
19. Choi, S., and Spiro, T. G. (1983) J . Am. Chem. SOC.105, 3683-3692
20. Riggs, A. (1981) Methods Enzymol. 131, 350410
21. Jeyarajah, S., and Kincaid, J. R. (1990) Biochemistry 29,5087-5094
22. Proniewicz, L. M., Paeng, I. R., and Nakamoto, K. (1991) J . Am. Chem. SOC.
113,3294-3303
23. Hu, S.(1993) Inorg. Chem. 32, 1081-1085
24. Kitagawa, T.,Ordries, M. R., Rousseau, D. L., Ikeda-Saito, M., and Yonetami,
T. (1982) Nature 298,869
25. Bruha, A,, and Kincaid, J . R. (1933) J . Am. Chem. SOC.110,6006
26. Proniewicz, L. M., and Kincaid, J. R. (1990) J. Am. Chem. SOC.112,675
27. Li, X.-Y., Czernuszewicz, R. S., Kincaid, J. R., and Spiro, T.G. (1989) J . Am.
Chem. Soc. 111, 7012-7023
28. Li, X.-Y., and Spiro, T. G. (1988) J . Am. Chem. SOC. 110, 6024-6033
29. Ray, G . B., Li, X.-Y., Ibers, J. A., Sessler, J. L., and Spiro, T. G. (1994) J . Am.
Chem. SOC.116, 162-176
30. Wells, A. V., Sage, F. T., Morikis, D., Champion, P. M., Chiu, M. L., and Sligar,
S. G. (1991) J. Am. Chem. Soc. 113,9655
S. Hirota, T. Ogura,E.H.Appelman, K. Shinzawa-Itoh, S.