Tải bản đầy đủ - 0 (trang)
A. Laser Photolysis: A Sledgehammer or a Scalpel?

A. Laser Photolysis: A Sledgehammer or a Scalpel?

Tải bản đầy đủ - 0trang

Infrared Studies in Heme Proteins


perturbation is expected to be large and sustained for up to several tens of

ps. Since that theoretical study, the electronic and thermal consequences

of light absorption have been probed experimentally with both near- and

mid-IR spectroscopy. The results of these experimental studies are now


1. Near-IR Study of Heme Relaxation

To experimentally probe the electronic and thermal consequences of flash

photolysis, a femtosecond time-resolved near-IR study of photoexcited Mb

was undertaken (22). This study probed the spectral evolution of band III,

a weak εmax ³ 100 M 1 Ð cm 1 near-IR charge transfer transition (14)

centered near 13, 110 cm 1 that is characteristic of five-coordinate ferrous

hemes in their ground electronic state S D 2 . Because band III is absent

when the heme is electronically excited, the dynamics of its reappearance

provides an incisive probe of relaxation back to the ground electronic state.

Moreover, because the spectral characteristics of band III (integrated area;

center frequency; line width) correlate strongly with temperature (23–26),

the spectral evolution of band III also probes its thermal relaxation.

Time-resolved absorbance spectra are typically recorded as difference

spectra with depletion of the ground state population appearing as negativegoing ground state features and the photoexcited population appearing

as positive-going features in absorbance and as negative-going features

in stimulated emission. For photolyzed Mb, no stimulated emission is

observed in the near-IR region, so the photoexcited population appears only

as positive-going features. Because the ground state bleach and the photoproduct absorbance have features in the same spectral region, the spectral

evolution of the photoexcited population cannot be measured without interference from the ground state bleach. Nevertheless, the photoproduct spectral evolution can be recovered from the transient absorption spectrum by

adding an appropriately scaled ground state absorbance spectrum. The scale

factor required is simply the fraction of Mb photoexcited by the pump pulse.

Whereas this fraction is not easily determined by direct measurement, it

can be determined by indirect measurement using a closely related system:

MbCO. The integrated area of band III at equilibrium was compared with

that of MbCO after photolysis under identical conditions (i.e., same heme

concentration and pump energy). Assuming that the quantum yield for the

photodissociation of MbCO is unity and correcting for small differences in

the absorbance of Mb and MbCO at the pump wavelength, the fraction of

photoexcited Mb within the probe-illuminated volume of the sample was

determined (22). The spectra shown in Fig. 4 were recovered by adding an

Copyright © 2001 by Taylor & Francis Group, LLC


Lim et al.

Figure 4 Time-resolved near-IR absorbance spectra of photoexcited Mb in

70/30 (w/w) glycerol/water at 1.00, 1.78, 3.16, 5.62, and 10 ps (open circles;

top to bottom). The spectra were recovered by adding an appropriately scaled

ground state Mb absorbance spectrum (dotted line) to the corresponding transient

near-IR absorbance spectra. The complete series of spectra (8 per decade) were

modeled (solid lines) with a temperature- and population-parameterized Gaussian

function plus a cubic polynomial background. The least-squares parameters for the

time-dependent population (B) and temperature rise (C) are plotted on a logarithmic

time scale. The data beyond 4 ps are well described (solid lines) by the functions

Population D 1 exp t/3.4 š 0.4 ps) and T D 140 K exp ( t/6.2 š 0.5 ps).

The dotted curves are extrapolations of these functions to early times, and the

dash-dot curves reflect the estimated uncertainty. (Adapted from Ref. 22.)

appropriately scaled equilibrium band III spectrum (12.8%; illustrated by

the dashed line) to each of the transient absorption spectra. Note that the

spectra have not been offset from one another, i.e., the differences in the

background are real.

Spectra at times earlier than 1 ps (not shown in Fig. 4) reveal a

1750 cm 1 broad (FWHM) feature centered at 12, 160 cm 1 that decays

with a 1.0 š 0.1 ps time constant (22). Because this band is red-shifted

more than 5, 500 cm 1 from the Q-band, it cannot be a vibronic band

Copyright © 2001 by Taylor & Francis Group, LLC

Infrared Studies in Heme Proteins


associated with the ground electronic state of a “hot” heme. The integrated

intensity of this feature is within a factor of two of the bleach of the Q-band,

suggesting that this relaxation pathway is likely the dominant pathway.

Finally, because this relaxation is much quicker than the reappearance of

band III, the 1 ps time constant corresponds to internal conversion between

two excited electronic states, not ground state recovery.

The spectral evolution in Fig. 4 reveals a broad, featureless

background offset that decays in amplitude as band III both grows in

integrated intensity and blue shifts in frequency. The growth of band III

arises from excited state relaxation back to the ground electronic state,

and its blue shift arises from cooling of the heme. To quantify this

spectral evolution, each time-resolved near-IR spectrum was modeled with

a Gaussian function added to a cubic polynomial background. The nonlinear

least-squares parameters characterizing band III (integrated area; center

frequency; FWHM) were found to evolve systematically beyond ¾3 ps but

varied less systematically at earlier times where the amplitude of band III

is small. To improve the analysis, an additional constraint was imposed

on the parameters describing band III. This constraint is based on the

fact that band III varies systematically with temperature. To implement

this constraint, the equilibrium band III spectrum was measured about

every 2 degrees from 0 to 70° C, and the temperature dependencies of its

integrated area, center frequency, and FWHM were modeled as quadratic

polynomials in temperature. Given this parameterization, a two-parameter

characterization of band III (population and temperature) was employed to

model the spectral evolution shown in Fig. 4 (22). This parameterization of

the Gaussian function required one less parameter and therefore provided

more robust estimates of those parameters. The experimental band III

spectra are well described by this model at all delay times shown in Fig. 4,

which also depicts the time dependence of the ground state population

and the heme temperature rise. At 4.22 ps, the temperature of the heme is

estimated to be approximately 100° C, a temperature modestly beyond the

range over which band III was characterized. To minimize errors arising

from extrapolating too far beyond reliable data, the time dependence of the

ground state population and the heme temperature were modeled with the

fitting range confined to times beyond 4 ps (solid lines). Beyond 4 ps, the

dynamics were well described by exponential decays with the ground state

recovering with a time constant of 3.4 š 0.4 ps and the thermal relaxation

proceeding with a time constant of 6.2 š 0.5 ps.

The near-IR study of band III suggests that relaxation back to the

ground electronic state proceeds with a 3.4 ps time constant. This value

Copyright © 2001 by Taylor & Francis Group, LLC


Lim et al.

is consistent with a 4.8 š 1.5 ps excited state lifetime determined using

saturated resonance Raman spectroscopy (27) and a 3.2 ps decay seen in

Soret band pump-probe studies (28). A recent Soret band pump-probe study

reported a 3.6 š 0.2 ps time constant for the ground state recovery and

strengthened that assignment by measuring the time-dependent absorption

anisotropy at 800 nm using diffractive optics-based heterodyned detection

of the imaginary component of the material nonlinear susceptibility (29).

The four different methods are all in excellent agreement with each other.

Because the near-IR study was carried out with Q-band excitation while the

other studies were carried out using Soret band excitation, it appears the

dynamics of ground state recovery are nominally independent of excitation


What are the consequences of the ¾3.4 ps electronic relaxation back

to the ground state? Ligands such as O2 , CO, and NO bind to heme in its

ground electronic state but are ejected when the heme becomes photoexcited. It is difficult to predict whether the intermediate excited electronic

states have a propensity to bind ligands. However, the lack of ps geminate

rebinding with CO suggests that these intermediate electronic states are

not receptive to ligands. Therefore, the electronic excitation may hold the

ligand at bay for a few ps, providing time for the surrounding protein to trap

the ligand in a docking site. It is worth noting that the intermediate excited

electronic states would have an influence on other ps time-resolved studies,

i.e., spectra measured at times shorter than a few ps would be complicated

by the presence of hemes in various electronic states.

The near-IR study of band III suggests that thermal relaxation of

photoexcited Mb is exponential with a 6.2 ps time constant. In reality,

this thermal relaxation is a nonequilibrium process, so any characterization of a time-dependent heme temperature is only approximate. Further

complicating matters, electronic relaxation proceeds through more than

one intermediate state so the photon energy is converted into heat in a

sequence of steps that can be separated by as much as a few picoseconds.

Consequently, the “temperature” of an individual photoexcited heme would

be expected to jump suddenly with each step of electronic relaxation and

would decay as the excess kinetic energy flows into the surroundings. The

time-resolved absorbance spectra in Fig. 4 probe an ensemble of hemes,

each having its own thermal history. In contrast, the parameterization of

band III used to model the spectra in Fig. 4 was made with the ensemble

of hemes in thermal equilibrium. Therefore, the calibration of temperature

in Fig. 4 is only approximate. Nevertheless, one might expect the near-IR

measurements to provide a reasonable ensemble-averaged estimate of the

Copyright © 2001 by Taylor & Francis Group, LLC

Infrared Studies in Heme Proteins


time-dependent heme temperature, especially for times beyond the electronic relaxation time of 3.4 ps. Therefore, the 6.2 ps thermal relaxation

time constant should be reasonably accurate. This conclusion is supported

by recent measurements of the real component of the material nonlinear

susceptibility of photoexcited Mb (29). They assigned a decay component of

5–7 ps to thermal relaxation of the heme, which is in very good agreement

with the 6.2 š 0.5 ps estimate from the near-IR band III study.

Mizutani and Kitagawa measured the time-dependent Stokes and antiStokes Raman intensities of the heme 4 band after photoexcitation and used

the relative intensities to estimate its “temperature” and thermal relaxation

dynamics (30). They found the population relaxation to occur biexponentially with 1.9 ps (93%) and 16 ps (7%) time constants. The dominant

1.9 ps population relaxation correlates with a 3.0 ps thermal relaxation,

which is a factor of 2 faster than the ensemble averaged temperature

relaxation deduced from the near-IR study of band III. The kinetic energy

retained within a photoexcited heme need not be distributed uniformly

among all the vibrational degrees of freedom, nor must the energy of all

vibrational modes decay at the same rate. Consequently, a 6.2 ps ensembleaveraged estimate of the heme thermal relaxation is not necessarily inconsistent with a 3 ps relaxation of 4 .

The near-IR measurements were performed on photoexcited Mb in

which all the photon energy is converted into heat. In contrast, when

photolyzing MbCO, a portion of the photon energy is required to break

the Fe–CO bond and the remainder is converted into heat. For example,

if a 555 nm photon was used to photolyze the sample (corresponding to

the center of the heme Q-band absorption) and if the bond dissociation

energy of Fe–CO was 16.2 kcal/mol (31), then approximately two thirds

of the photon energy would be available to heat the heme. Therefore, the

magnitude of the heating effect in photolysis studies of MbCO would be

only two thirds as great as that reported for photoexcited Mb. Because

the heme cooling rate should be equally fast in photolyzed MbCO, the

protein quickly loses memory of how the ligand was dissociated in the first

place. Consequently, after the first 10–20 ps, the ligand dynamics should

be independent of the detachment mechanism and photolysis studies of

ligand rebinding and escape dynamics should be physiologically relevant.

2. Mid-IR Study of CO Relaxation

Whereas near-IR spectroscopy provides an incisive probe of the electronic

and thermal state of the heme after flash photolysis, it tells us nothing about

the state in which the ligand is created. Mid-IR spectroscopy, on the other

Copyright © 2001 by Taylor & Francis Group, LLC


Lim et al.

hand, can probe the vibrational state of the dissociated ligand, provided it is

not a homonuclear diatomic. The impact of flash photolysis on a dissociated

ligand was recently investigated with time-resolved IR spectroscopy (32).

A sample containing Hb13 CO was photolyzed with picosecond laser pulses

(35 ps; 527 nm) and probed with optically delayed femtosecond IR pulses

(200 fs; ắ5 àm). The photolysis pulse dissociated approximately 35% of

the Hb13 CO within the probe-illuminated volume and time-resolved mid-IR

absorbance spectra were measured at a series of times that were equally

spaced on a logarithmic time scale (see Fig. 5). Four features are readily

apparent in the early time spectra, but only two features, labeled B1 and

B2 [following Alben et al. (33)], survive in the later time spectra. The

surviving features correspond to 13 CO in its ground vibrational state D 0

and trapped within a protein docking site (34). The two satellite features

labeled BŁ1 and BŁ2 decay exponentially with a 600 š 150 ps time constant.

BŁ1 and BŁ2 are well described as red-shifted replicas of B1 and B2 , and the

experimentally determined shift between B1 B2 and BŁ1 BŁ2 is 26 cm 1 ,

0 and hot band

similar to the 25.3 cm 1 shift between the ground 1


1 vibrational transitions of 13 CO in the gas phase (35). Therefore,

the decaying features are assigned to 13 CO generated in its first excited

vibrational state D 1 .

The 600 ps vibrational cooling time constant for “docked” CO is far

slower than the ¾18 ps cooling time constant measured for bound CO in

MbCO (A1 -state) (36,37) and in HbCO (37) and is far slower than the

6.2 ps time constant for cooling of the heme (22). The sluggish cooling

rate of photodetached CO demonstrates that coupling between the highfrequency CO vibrational motion and the lower-frequency acceptor modes

of the heme and protein is weak. A theoretical study of the vibrational

relaxation rate has been performed using the Landau-Teller model, and

excellent agreement between experiment and theory was obtained (32).

Moreover, that study included a normal mode analysis which identified

those protein residues that act as the primary “doorway” modes in the

vibrational relaxation of the oscillator. From that study, they concluded

that the distal histidine plays an important role in the vibrational


From the data in Fig. 5 one can also deduce the vibrational temperature of photodetached CO. Assuming the absorbance cross section for the

first hot band transition 2

1 is twice that for the ground state transition 1

0 (38), the population of CO produced in its first excited

vibrational state represents only 3.6% of the total photolyzed population.

A 3.6% nascent yield in D 1 corresponds to a Boltzmann temperature

Copyright © 2001 by Taylor & Francis Group, LLC

Infrared Studies in Heme Proteins


Figure 5 (A) Picosecond time-resolved mid-IR absorbance spectra of photolyzed

Hb13 CO. Photolysis of Hb13 CO detaches 13 CO from the heme, whereupon it

becomes trapped in a docking site located near the heme-binding site. The


docked 13 CO is produced predominately in its ground vibrational state

and gives rise to two major features, B1 and B2 . A small (3.6%) portion of

D 1 and gives rise to

the 13 CO is produced in its excited vibrational state

two satellite features, BŁ1 and BŁ2 . Vibrational relaxation back to the ground state

D 1 ! 0 causes the satellite features to disappear. The solid curves were

obtained by least-squares fitting the experimental spectra with a constrained model

involving Gaussian functions. (B) The time-dependent population of CO in its D 1

vibrational state is also shown (reported as a percentage of the total population of

“docked” CO). The nascent yield of vibrationally excited CO is about 3.6%, and this

population decays exponentially with a time constant of 600 š 150 ps. (Adapted

from Ref. 32.)

of 825 K, which is hotter than the Ä478 K expected if the photon energy

beyond that required to break the Fe–CO bond were distributed uniformly

over all degrees of freedom of the heme-CO (21,22,39). Clearly, CO is not

in thermal equilibrium with the heme when it is ejected. This result should

come as no surprise because the time scale for CO photodetachment is

within 100 fs of photoexcitation (28,34), a time too short for the CO to

emerge in thermal equilibrium with the heme. Rather, the excited vibrational state population is dictated by the dissociation mechanism, which

Copyright © 2001 by Taylor & Francis Group, LLC


Lim et al.

strongly favors production of ground state CO. This, too, should come as

no surprise: one might expect the dominant pathway for generating vibrationally excited CO to be the impulsive half collision with Fe after the

electronic potential becomes repulsive. Because the CO bond lengths in

bound and unbound CO are similar, energy transfer through an impulsive

half collision would be rendered inefficient (40,41). Because the nascent

population of vibrationally excited CO is small, it is of little consequence

to measurements of ligand rebinding and escape dynamics.

B. Evidence for a Ligand Docking Site in the Heme Pockets of

Mb and Hb

Time-resolved mid-IR spectra of several photolyzed heme proteins are

shown in Fig. 6. The negative-going features correspond to loss of bound

CO (A states) and the positive-going features correspond to CO dissociated from the heme (B states). Note the transition frequencies and relative

intensities of the A and B states. When CO is bound to heme, back-bonding

between the CO -orbitals and the iron d-orbitals weakens the CO bond and

enhances its transition moment (42). Compared to “free” CO, the bound

CO vibrational frequency is red shifted about 200 cm 1 and its integrated

oscillator strength at 5.5 K is enhanced 21.7 š 1.6 times (33).

The integrated areas, center frequencies, and line widths of the A

and B states in Fig. 6 were characterized by modeling these features as a

sum of Gaussians on a quadratic (B states) or cubic (A states) polynomial

background (11). The center frequencies and line widths of the A states

in sw -Mb13 CO: D2 O, which were recorded with ¾200 fs duration mid-IR

pulses, were found to be comparable to those reported previously (33,43).

The agreement between the ultrafast time-resolved and static measurements

of the A states should not be surprising: the time-resolved A state spectra

measured at 100 ps simply recover the equilibrium A state spectra plotted

in a negative-going fashion. On the other hand, the B states near ambient

temperature are short-lived intermediates whose characterization requires

time-resolved methods. The B state spectra shown in Fig. 6B reveal features

that are significantly narrower than any feature ever assigned to non bonded

CO in the condensed phase near room temperature. For example, spectra of

CO dissolved in a host of organic solvents (44,45) are very similar to the

experimental CO spectrum in cyclohexane shown in Fig. 3A, the FWHM

of which is 98 cm 1 . In contrast, the B states in Fig. 6B are more than

10 times narrower. Clearly, the interior of the protein, which is composed

primarily of hydrophobic residues, presents an environment quite different

from that of liquid organic solvents. Moreover, this unique environment

Copyright © 2001 by Taylor & Francis Group, LLC

Infrared Studies in Heme Proteins


Figure 6 Time-resolved mid-IR spectra of photolyzed h-Mb13 CO:D2 O (gray

line), sw -Mb13 CO:D2 O (filled circles), and Hb13 CO:D2 O (black line). The spectra

were recorded at 100 ps and 283 K with the A-state region (A) and B -state region

(B) collected independently. The A- and B -state designations follow the convention

of Alben et al. (33). To facilitate comparison among the spectra, the constant and

linear contributions to the polynomial fit of the background have been subtracted

from the measured spectra.

Copyright © 2001 by Taylor & Francis Group, LLC

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

A. Laser Photolysis: A Sledgehammer or a Scalpel?

Tải bản đầy đủ ngay(0 tr)