Role of Local Structure and Dynamics of Small Ligand Migration inProteins: A Study of a Mutated Truncated Hemoprotein fromThermobif ida fusca by Time Resolved MIR SpectroscopyBarbara Patrizi,†,§ Andrea Lapini,†,‡,§ Mariangela Di Donato,*,†,‡,§ Agnese Marcelli,† Manuela Lima,†,§

Roberto Righini,†,‡,§ Paolo Foggi,†,§,∥ Paola Baiocco,# Alessandra Bonamore,⊥ and Alberto Boffi⊥

†LENS (European Laboratory for Nonlinear Spectroscopy) Via N. Carrara 1, Sesto Fiorentino, Florence 50019, Italy‡Dipartimento di Chimica, Universita ̀ di Firenze, via della Lastruccia 3-13, Sesto Fiorentino, Florence 50019, Italy§INO (Istituto Nazionale di Ottica), Largo Fermi 6, Firenze, Florence 50125, Italy∥Dipartimento di Chimica, Universita ̀ di Perugia, via Elce di Sotto 8, Perugia, Umbria 06123 Italy⊥Istituto Pasteur, Fondazione Cenci Bolognetti c/o Dipartimento di Scienze Biochimiche, Universita ̀ “La Sapienza”, piazzale AldoMoro 5, Rome, Rome 00185, Italy#Istituto Italiano di Tecnologia, Center of Nano Life Sciences, Viale Regina Elena 291, Rome, Rome 00161 Italy

*S Supporting Information

ABSTRACT: Carbon monoxide recombination dynamics in a mutant of the truncatedhemoglobin from Thermobida fusca (3F-Tf-trHb) has been analyzed by means ofultrafast Visible-pump/MidIR-probe spectroscopy and compared with that of the wild-type protein. In 3F-Tf-trHb, three topologically relevant amino acids, responsible for theligand stabilization through the formation of a H-bond network (TyrB10 TyrCD1 andTrpG8), have been replaced by Phe residues. X-ray diffraction data show that Pheresidues in positions B10 and G8 maintain the same rotameric arrangements as Tyr andTrp in the wild-type protein, while Phe in position CD1 displays significant rotamericheterogeneity. Photodissociation of the ligand has been induced by exciting the samplewith 550 nm pump pulses and the CO rebinding has been monitored in two mid-IRregions respectively corresponding to the ν(CO) stretching vibration of the iron-boundCO (1880−1980 cm−1) and of the dissociated free CO (2050−2200 cm−1). In both themutant and wild-type protein, a significant amount of geminate CO rebinding isobserved on a subnanosecond time scale. Despite the absence of the distal pockethydrogen-bonding network, the kinetics of geminate rebinding in 3F-Tf-trHb is very similar to the wild-type, showing how thereactivity of dissociated CO toward the heme is primarily regulated by the effective volume and flexibility of the distal pocket andby caging effects exerted on the free CO on the analyzed time scale.

1. INTRODUCTION

Heme proteins are widely distributed in nature where theycover multiple functions, related to the transport and storage ofdifferent diatomic ligands, such as O2, NO, and CO. Thedynamics of ligand binding and escape are tightly connected tothe structural properties of the hosting protein. Structuralconstraints play a pivotal role both at a local level, byestablishing specific interactions between the topologicallyrelevant amino acid residues surrounding the heme ligand, andon a more global level through the structural cavities andchannels that pave the way to ligand income and escape to andfrom the protein matrix.1−4

Since the photodissociation of bimolecular ligands can beinduced by a short laser pulse, the dynamics of recombinationand escape of O2, NO, and CO in several heme containingproteins such as myoglobin, hemoglobin and peroxidases5,6 hasbeen extensively studied using transient spectroscopy techni-ques. Given its strong affinity for most heme proteins, CO has

often been used as a probe to investigate the dynamics ofphotodissociation and rebinding.1,7−9 In particular, the highsensitivity of the ν(CO) stretching vibration of the heme−COcomplex toward the electric field generated by the proteinenvironment offers a unique possibility to follow COrecombination dynamics in the infrared spectral window.Monitoring the transient infrared spectrum after laser pulsephotolysis thus provides specific pieces of information aboutthe influence of structural and electrostatic properties of thedistal heme pocket on the ligand dissociation and rebindingprocesses. Time resolved spectroscopy in the visible region issensitive to the dynamics of the electronic state of the hemecofactor, thus reflecting the overall ligand recombinationprocess, whereas measurements in the infrared region are

Received: May 7, 2014Revised: July 10, 2014Published: July 14, 2014

Article

pubs.acs.org/JPCB

© 2014 American Chemical Society 9209 dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−9217

specific of the ligand migration properties. As such, the directprobe of the vibrational bands of the ligand before and afterphotolysis, provides a direct snapshot of the fate of photolyzedCO molecule within the protein matrix. This sensitivity isfurther enhanced by the possibility of comparing different sitedirected mutants of a heme-protein, thus providing a detailedunderstanding of the specific interactions between the aminoacids contained in the heme pocket and the CO ligand and oftheir effect on the rebinding dynamics.In this work, we have investigated the recombination

dynamics of the photodissociated CO in the triple mutant ofa truncated hemoglobin protein, Thermobif ida fusca (Tf-trHb),by using Visible-pump/MidIR-probe spectroscopy.10 Therebinding properties of the parent wild type11 protein havebeen characterized previously1,2 thus allowing a direct under-standing of the influence of the mutated residues on the liganddynamics and related protein contributions.Truncated Hemoglobins (trHbs) are a family of small oxygen

binding proteins widely distributed among bacteria, plants, andprotozoa.12,13 They are characterized by a common generalpattern of the heme pocket, which contains an ensemble ofpolar residues capable of forming hydrogen bonds with theiron-bound ligand, usually defined as “ligand inclusive hydrogenbonding network”. It has been suggested that TrpG8 plays animportant role in ligand binding and stabilization, althoughother amino acids in topological positions E7, E11, CD1, andB10 can drastically modify the interaction network.3,14,15 Thecrystal structure of Tf-trHb16 shows that the heme distal pocketinvolves, beside TrpG8, two Tyr residues, TyrB10 andTyrCD1, which are in close proximity with the iron-boundligand. Furthermore, it has been observed that the protein canadopt two different conformations, differing by the number ofhydrogen bonds formed with the iron-coordinated COmolecule.3,17,18 In the first conformation, both TrpG8 andTyrCD1 are hydrogen bonded to the iron-coordinated CO,whereas in the second one, only the H-bond with TrpG8 ismaintained. In addition, a water molecule, which is hydrogenbonded to TyrCD1, when the latter residue is not directlycoordinated with the iron ligand, has been proven to befundamental in determining the structural properties of thedistal pocket.4,15 In this work, the ligand dynamics of a triplemutant (3F-Tf-trHb) in which the three CO-interacting aminoacid residues, TrpG8, TyrB10, and TyrCD1 are replaced byPhe residues (see Figure 1), has been studied. The triplemutation actually destroys the hydrogen bonding network

around the ligand, and makes the distal pocket substantially lesspolar than in the WT protein. In previous studies,1,2,4 it hasbeen shown that CO recombination dynamics in the WT Tf-trHb is significantly faster if compared with other hemoglobinor myoglobin proteins, and that a substantial amount of COrecombines via a monophasic geminate process on thenanosecond time scale. The occurring of such a rapid geminaterecombination has been ascribed to the particular structure ofthe ligand environment, which acts as a cage around thedissociated CO, preventing its escape toward the solvent, inanalogy with what observed also in other truncatedhemoglobins.1,3,19,20 Another factor which could contribute tothe observed fast rebinding is the possibility that in truncatedhemoglobins the primary docking site of the dissociated CO iscloser to the heme iron as compared with myoglobin, whichwould favor a faster rebinding. Such a situation has beenobserved in case of human hemoglobin, through a recent time-resolved X-ray study.21

In this work, the direct comparison of the 3F-Tf-trHb andthe WT has allowed to establish how the specific interactionsinside the distal pocket influence the CO rebinding:surprisingly, as also observed in the truncated hemoglobin ofMycobacterium tubercolosis19 we found that on a shortnanosecond time scale the binding properties and therecombination dynamics are substantially robust againstmutations, since only little difference is observed comparedto the WT. However, when considering a longer time scale,spanning up to the millisecond regime, notable differences areobserved.4 All the experimental results can be rationalized inthe frame of a recently proposed kinetic scheme, whichaccounts for the presence of several cavities inside the protein,where the dissociated CO is temporary located before eitherrecombining with the heme or escaping toward the solvent.4

2. MATERIALS AND METHODS2.1. Protein Expression and Purification-Sample

Preparation. Cloning, expression, and purification of triplemutant 3F-Tf-trHb (Tyr45Phe, Tyr67Phe, Trp119Phe) ofacidic surface variant (ASV) hemoglobin from Thermobif idafusca (Tf-trHb) (Arg91Glu, Phe107Glu) were carried out aspreviously reported.3

Protein solution for Vis-pump Mid-IR-probe measurementswas prepared by dissolving the samples, obtained as lyophilizedprotein, in a TRIS-HCl buffer 0.2 M in D2O (pD = 8). Theconcentration of the protein solution used for time-resolvedmeasurements was about 10 mM. Reduction of the protein wasaccomplished by adding a freshly prepared anaerobic solutionof sodium dithionite in stoichiometric excess to the proteinsolution, previously degassed with nitrogen. Carbon monoxide(Rivoira), was gurgled at low flux intensity, and the sealedprotein solution was saturated with 1 atm CO for 15 min. Inthis way CO is homogeneously distributed in spite of the highviscosity of the sample. Samples for transient infraredmeasurements were prepared by squeezing about 40 μL ofsolution between two calcium fluoride windows (3 mmthickness) separated by a 50 μm Teflon spacer. The OD atthe excitation wavelength was about 0.6.

2.2. X-ray structure determination. 2.2.1. Crystallization.Crystals were obtained at 21 °C in 3 days by vapor diffusionusing the hanging drop technique. Drops were prepared bymixing 1 μL of protein (15 mg/mL) equilibrated against 50mM of phosphate buffer, pH 7.0 and 1 μL of reservoir solution(2.7 M NaCl, 0.1 M potassium acetate pH 6.0). Crystals

Figure 1. Close up view of 3F-Tf-trHb distal heme pocket. The picturehas been generated with UCSF Chimera,22 the pdb file is underdeposition.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179210

appeared as very long and thin rods. In order to optimize theman additive screen was performed in the crystallizationconditions. Crystals were truncated prisms of dimensionsabout 0.3 × 0.2 × 0.1 mm3 and were obtained by adding 5 mMEDTA as final concentration in the drop. Crystals were cryo-protected in a solution containing 75% v/v of the reservoirsolution and 25% v/v glycerol and were flash-frozen in liquidnitrogen.2.2.2. Data Collection and Data Reduction. Single

wavelength data set (λ = 0.893 Å) was collected at thebeamline BL 14−1 at the synchrotron radiation source BeSSYIIin Berlin, Germany, using a CCD detector at a temperature of100 K. The data set was processed with DENZO23 and scaledwith SCALEPACK.23 The crystals belong to the P4122 spacegroup with the following cell dimensions: a = 78.37 Å, b =78.37 Å, c = 89.80 Å. The data scaling gave a R-merge of 0.128for 3954 unique reflections. Crystal parameters and completedata-collection statistics are given in Table S1 of the SupportingInformation (SI). The value of VM calculated using the CCP4suite24 (1994) is 3.93 Å3Da−1 with only one molecule in theasymmetric unit. This indicates that the crystals display a veryhigh solvent content (68.7%).25

2.2.3. Structures Solution, Model Building and Refine-ment. The structure of the triple mutant of the ASV-hemoglobin from Thermobif ida fusca was solved by molecularreplacement using wild type form as search model (pdb code2BMM). Refinement was performed using the maximumlikelihood method with the program REFMAC526 and modelbuilding was carried out using the program COOT11 (see TableS1 in the SI). The final Rcrys for all resolution shells (50−3.4Å) calculated using the working set reflections is 27,2%, and theFree R value calculated using the test set reflections is 31,2%.The final model contains 132 residues (31−160 residues), 1heme molecule, and 1 acetate anion. The quality of the modelwas assessed using COOT and PROCHECK.11 The atomiccoordinates and structure factors have been deposited in theProtein Data Bank (pdb code: under deposition).2.3. Visible-Pump/MidIR-Probe Spectroscopy. Meas-

urements were performed probing both the absorption regionof the ν(CO) stretching vibration of the iron-bound CO(1880−1990 cm−1) and the dissociated free CO absorption(2050−2200 cm−1). CO dissociation was induced by excitingthe heme with a 550 nm laser pulse. The experimental setup forthe infrared differential absorption measurements has beendescribed in detail in reference.10 Briefly, a Ti/sapphireoscillator/regenerative amplifier, operating at 1 kHz, (Legend,Coherent) was used to pump a home-built optical parametricgenerator and amplifier with difference frequency generation,which produced a tunable output (2.5−10 μm) with a spectralwidth of ∼200 cm−1. The output of a HgCdTe camera system,placed behind a spectrograph, was read out every shot at arepetition rate of 1 kHz and a sampling resolution of 4 cm−1.Excitation pulses at 550 nm (energy = 200 nJ) were obtainedby sum frequency generation of the idler output of acommercial optical parametric generator (TOPAS, LightConversion) with a portion of the fundamental output at 800nm. A moveable delay line made it possible to increase thetime-of-arrival-difference of the pump and probe beams up to1.8 ns. The pump beam polarization was set to magic anglewith respect to the probe beam by rotating a λ\2 plate. Thesample was moved with a home-built scanner to refresh thesolution and avoid photodegradation. Home-written softwarewas used to collect the data over the two different spectral

windows respectively between 1880 and 1975 cm−1 and 2050−2200 cm−1. Every window was recorded with a freshly preparedsample and measured at least three times. In order to obtain agood signal-to-noise ratio in the case of the free CO signal,which has a small absorption cross section, 10−15 data sets,each recorded accumulating 2000 laser shots per time point,were collected and averaged. The integrity of the sample hasbeen checked by FTIR (Bruker Alpha-T) and visible absorption(PerkinElmer LAMBDA 950) before and after the time-resolved measurements.

2.3.1. Data Analysis. For the quantitative analysis of thetime-resolved spectral data we used a combined approach,consisting of singular values decomposition (SVD) and thesimultaneous fitting of all the collected kinetic traces (globalanalysis). First, we extrapolated the number of componentsusing SVD,27−29 then we analyzed the whole ensemble ofkinetic data by means of a global fitting procedure. Thecombination of global analysis and SVD is a very helpfulanalysis protocol, as it provides a good control on the numberof components used to fit the data.30−35 The time constantsresulting from the fitting of the right singular vectors have beenused as the starting point for the subsequent global analysis.The aim of global analysis is to decompose the two way datamatrix into time-independent spectra and wavelength in-dependent kinetics.34−37 Once the number of componentshas been identified, the second step involves the para-metrization of the time evolution of the relative intensities ofthe spectral components. This was accomplished by assuming afirst-order kinetics, describing the overall temporal evolution asthe sum or combination of exponential functions. Globalanalysis was performed using the GLOTARAN package(http://glotaran.org).35,38−41 We employed a linear unidirec-tional “sequential” model. In discussing our results, the spectraassociated with the various time constants are termed EvolutionAssociated Dif ference Spectra (EADS).

3. RESULTS3.1. Three-Dimensional Structure. The three-dimen-

sional structure of the triple mutant of ASV-Tf-trHb has beendetermined at 3.3 Å resolution. The superposition between themain-chain atoms of Tf-trHb and triple mutant ASV-Tf-trHbyields a root-mean-square deviation of 0.73 Å. The overallprotein structure is not affected by the acidic mutation ofPhe107Glu and Arg91Glu as previously described.3 Bothmutations remain solvent exposed and Phe107Glu mutationdoes not alter the conformation of the loop containing theproximal histidine, namely His106-Phe8, which is 2.20 Å distantfrom ferric ion. The structure contains additional residues at N-terminal (residues 31−33) and at C-terminal tails (residues157−160) and a very flexible loop involving residues from 68 to74. B-average analysis shows a very high mean value which is103.58 Å2 for the main chain, and it was impossible to buildwater molecules in the electron density due to the lowresolution of the data. Ramachandran plot analysis24 revealedtwo outliers residues (namely Pro155 and Asn159) which havebeen tolerated since belonging to C-terminus region. Theproximal heme pocket shows no substantial rearrangements.The His106-Phe8 is bonded to the Fe(III) of the heme with alonger distance than wild type (NFe(III) = 2.20 Å). Thedistal heme pocket is characterized by three mutations intophenylalanine of the ligand inclusive hydrogen bond network,namely TyrB10Phe-TyrCD1Phe-TrpG8Phe, (Tyr54Phe-Tyr67Phe-Trp119Phe) that abolish polar interaction with the

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179211

iron-bound ligand coordination shell. The heme ring is onlystabilized by interactions involving the propionyl groups withresidues R78 (NH1O1A = 3.0 Å), R97 (NH1O1D = 2.8Å) and Y93 (OHO2D = 2.5 Å).The acetate ion is almost parallel to the heme ring. The OXT

ion of acetate ion is 2.68 Å and O atom is 2.82 Å distant fromferric ion with an occupancy of 75% and a B factor = 91 Å2.Phe119-G8 replacing Trp119-G8 residue in the wild type11 isessentially parallel to the heme plane as Trp119. The regioncontaining Phe67 replacing Tyr67-CD1 is characterized by avery poor electron density with a very high B factor values. Thisloop (residues 67 to 74) has a different conformation respect tothe wild type and has been completely rebuilt.Electron density contour map within the active site of 3F-Tf-

trHb is shown in Figure 2. The map illustrates the contour

surface at the heme site bringing out evidence that the PheCD1occupancy is not restricted to the canonical position with thearomatic ring parallel to the heme plane but also broadlypopulate an outward position with the aromatic ring in skewedposition with respect to the heme plane. The distal ligandregion is also most likely populated by two conformers of theresident acetate molecule. Positive electron density is alsovisible on top of the ligand and might be attributed to watermolecules. However, the current resolution does not allowrefinement of such molecules.3.2. Steady State Spectra. The UV−vis and the FT-IR

spectra of 3F-Tf -trHb-CO complex are reported in Figure 3. Incontrast to the wild type protein,1,17 the FT-IR spectrum of the3F-Tf -trHb−CO complex displays a single band, centered at

1955 cm−1, indicating a single conformation for the heme−COadduct. The frequency of the bound-CO vibration is higherthan those measured in the WT protein (1920 and 1940 cm−1),as no hydrogen bonds, capable to stabilize the iron-bound CO,are formed.

3.3. Time Resolved Mid-IR Spectra. In order toinvestigate the CO recombination dynamics in 3F-Tf-trHbwith Vis-pump/MidIR-probe we have triggered ligand dis-sociation with 550 nm visible pump pulses. We have monitoredCO rebinding in two mid-IR regions: the region of the ν(CO)stretching vibration of the iron-bound CO (1880−1980 cm−1)and the dissociated free CO absorption region (2050−2200cm−1).

3.3.1. Bound CO. Time resolved spectra of 3F-Tf -trHb inthe bleaching region are reported in Figure S1 in the SI.Following CO photodissociation, only one bleaching band isobserved at 1955 cm−1, whose position well correspond withthe measured FT-IR spectrum of the CO adduct for thisprotein (Figure 4). A global analysis of the kinetic traces, using

a sequential kinetic scheme, reveals a biphasic recoveryoccurring within a nanosecond time scale. The kinetic tracetaken at the maximum of the bleaching band and the EADSobtained by global analysis are reported in Figure 4.The kinetic traces can be satisfactorily fitted with two

components, whose time constants are 250 ps and 1.4 ns,respectively. This kinetic behavior is similar to that previouslyobserved for the WT protein1 but, in this case, the fastpicosecond component seems to have a minor amplitude ifcompared with the WT protein (30% vs 40%). Thisobservation will be discussed in detail in the next section.

3.3.2. Dissociated CO. Measurements were carried out bysetting the excitation pulse at 550 nm, probing the stretchingregion for free CO absorption from 2050 to 2200 cm−1.Spectra in Figure 5 have been corrected for the presence of a

baseline due to substantial water absorption in this spectral

Figure 2. Contour map of electron density within the distal hemepocket for the 3F-Tf-Hb mutant. The map has been obtained at a 3.3Å resolution and displays an anomalously broad density on top of theheme iron as well as in the topological position occupied by PheCD1(Phe119).

Figure 3. UV−vis (left panel) and FT-IR (right pane, red line) spectraof the CO adduct of 3F-Tf-trHb. The FT-IR spectrum of the WTprotein is shown for comparison in the right panel (black line).

Figure 4. Kinetic trace (scattered points) at 1955 cm−1 with the fit(left panel) and the EADS (right panel) obtained by global analysis.

Figure 5. Time resolved spectra in the free CO region recorded atdifferent time delays after excitation of the sample with a 550 nm laserpulse.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179212

region, by subtracting a third order polynomial fit whichexcludes the 2100−2130 cm−1 region, where the signal isobserved. The details of this baseline subtraction procedure arereported in the SI. The fwhm average value found for the freeCO absorption band of the mutant protein is about 15 cm−1,about half the value with respect to the WT protein (∼ 25cm−1).1 In Figure 6, the kinetic trace measured at the

absorption maximum of the free CO signal is shown andcompared with that of the 1955 cm−1 bleaching band. The twokinetic traces have been scaled to be appropriately comparedand are in good agreement, confirming that the both the 250 psand the 1.4 ns time constants are due to fast geminaterecombination.In order to inspect if the bandwidth changes in time, we

estimated the temporal evolution of the integrated area of theCO band, and we compared it with the kinetic trace at 2120cm−1, corresponding to the band maximum. The two traces arein good agreement, thus indicating no significant bandwidthvariation as a function of time. (see Figure S2 in the SI).3.4. Comparison between Tf-trHb and 3F-Tf-trHb.

Although the kinetic constant for the CO recombination in theWT protein and 3F mutant are very similar (on the order of250 ps and 1.4 ns) some differences can be noted. In order tohighlight them, we have compared in Figure 7 the kinetic traces

measured for the two bleaching bands of Tf-trHb (1940 and1920 cm−1, kinetic data have been previously published by theauthors of this manuscript1) with the kinetic trace correspond-ing to the bleaching band observed for 3F-Tf -trHb (1955cm−1). The curves have been normalized on the long (ns) timescale. In the case of WT, two bleaching bands were observed inthe Mid-IR spectrum of Tf-trHb and related to two populations

of bound-CO.17 The reported comparison brings out clearly afaster decay in the case of 1940 cm−1 WT time trace withrespect to that at 1955 cm−1 of 3F-Tf -trHb. In contrast, thekinetic trace taken at 1920 cm−1 for WT displays a more similarkinetic behavior with respect to that of the mutant.As concerns the free CO region, a comparison between the

kinetic traces corresponding to the maximum of the free COabsorption bands for Tf -trHb and 3F-Tf -trHb is reported inFigure 8.

The dynamic evolution of the two signals is similar, except inthe first 30 ps, where a rise component is observed for the WTprotein. We have previously attributed this 30 ps component toa significant protein rearrangement involving the rotation orreorientation of one or more amino acid side chains in the COdocking site.1 This component was neither observed in the 3Fmutant analyzed in this study, nor for the truncatedhemoglobin from Bacillus subtilis (Bs-trHb), which was alsopreviously studied.1 The lack of the rise component can becorrelated with a major flexibility of the distal pocket of the WTTf-trHb, compared with the two other analyzed cases, whichtranslates in a higher structural rearrangement upon COphotodissociation. Note that the WT protein differ from boththe triple mutant and Bacillus subtilus in that it has a structuralwater molecule in the distal pocket, which may create a moreheterogeneous environment for the dissociated CO molecule.In 3F-Tf -trHb the environment within the distal heme pocketis less polar and apparently more homogeneous than that of theWT protein. A lower degree of protein rearrangement inducedby the ligand dissociation could explain the absence of a risecomponent in the kinetics of the photodissociated CO in 3F-Tf-trHb and Bs-trHb.

3.5. Free CO Band Shape as a Function of ChemicalEnvironment: A Qualitative Comparison. As a finalanalysis intended to understand the influence of the proteinenvironment on the dissociated CO and consequently of itscapacity to act as a probe on the local protein structure, we havecompared the shape of the free CO absorption band of 3F-Tf-trHb with those previously characterized for the WT proteinand the truncated hemoglobin Bs-trHb. The three spectra,recorded at 30 ps delay time after photodissociation, have beennormalized to the same intensity and reported in Figure 9.By modeling the curves with a Gaussian function, we have

found a fwhm of ∼27 cm−1 for Bs-trHb and 25 cm−1 for Tf-trHb while for triple mutant the value is almost halved (∼15cm−1). The CO absorption bands for Tf -trHb and 3F-Tf -trHbare centered at ∼2120 cm−1, while for Bs-trHb, the central

Figure 6. Kinetic trace at 2120 cm−1 (red line), corresponding to themaximum absorption in the free CO region in 3F-Tf -trHb. The tracehas been superimposed to the kinetic trace at 1955 cm−1 (black line),corresponding to the maximum absorption of the coordinated CO.

Figure 7. Comparison between the kinetic traces in the bleachingregion for Tf-trHb and 3F-Tf -trHb upon excitation at 550 nm. Tracesat 1940 cm−1 (red line) and 1920 cm−1 (black line) of Tf-trHb arecompared to the trace at 1955 cm−1 of 3F-Tf -trHb (green line). Thethree kinetic traces have been normalized on the long time scale.

Figure 8. Kinetic trace at 2120 cm−1 recorded at the maximumabsorption in the free CO region in Tf-trHb (red line) compared withthe corresponding kinetic trace measured in 3F-Tf -trHb at 2120 cm−1

(blue line) in the first 300 ps.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179213

frequency is located at ∼2130 cm−1. The shifts relative to thegas frequency of CO (ν0 = 2143 cm−1), reported in the SI, are13 cm−1 in the case of Bs-trHb and 23 cm−1 for the other twoproteins. In order to rationalize these observations, it is usefulto remind that in the case of horse myoglobin three free CObands, indicated as B bands have been characterized:42,43 B1,B′2 and B″2 respectively peaking at 2132, 2122, and 2117 cm−1.The high frequency (2132 cm−1) B1 state was assigned to theconformer with the O atom pointing back toward the hemeiron atom while the low frequency (2122 cm−1) B2 state to theopposite orientation. The red shifts from the value of 2143cm−1 for the gaseous CO implies weak interactions between theCO in the pocket and the residues constituting the distal hemepocket. The magnitude of the shift increases with bindingenthalpy.44 The 2130 cm−1 of Bs-trHb resembles the B1 band ofmyoglobin and indicates a strong interaction between thedissociated CO and the distal residues. Tf-trHb and its 3Fmutant exhibit a free CO stretching band around 2120 cm−1

and the shift relative to the gaseous CO is two times largerrespect to Bs-trHb. In Tf -trHb, the distal heme pocket is verysimilar to that of Bs-trHb, except for the presence of a watermolecule observed in Tf-trHb. This water molecule, asmentioned before, could diminish the volume of the hemepocket and could form a narrower cage for the free COmolecule, although, due to its mobility and the possibility todynamically reorient or rearrange its position, increases theflexibility of the pocket. In the case of 3F-Tf -trHb, we observe avery narrow absorption band with respect to the WT protein.The substitution of the three polar amino acids TrpG8,TyrB10, and TyrCD1 of the WT protein with Phe residuesleads to a less-polar distal pocket which is apparently morehomogeneous than in WT, resulting in the narrower band. Inthis view, the possible displacement of the structural watermolecule in the mutant, due to the lack of hydrogen bondingresidues, results in a more ordered and less flexible distal pocketfor which the band narrowness indicates that the CO is in amore constrained environment. Finally, we notice that the freeCO absorption bands within all the analyzed proteins arenarrower than that of CO dissolved in organic solvents (seeFigure S4 in the SI section); the shifts relative to the gaseousCO central frequency are larger for the CO docked in theprotein matrix respect to those observed for the CO dissolvedin pure solvents. This can be either ascribed to specificinteractions and/or to high values of viscosity within theprotein matrix.

4. DISCUSSION

In this study, we have analyzed the CO rebinding dynamics inthe triple mutant of the truncated hemoglobin Thermobidafusca where the three topologically most relevant amino acids,responsible for the ligand stabilization through the formation ofa H-bond network (TyrB10 TyrCD1, and TrpG8), have beenreplaced by Phe residues. The X-ray diffraction data of thetriple mutant, obtained at 3.3 Å resolution, indicate that the Pheresidues in positions B10 and G8 do maintain the samerotameric arrangements of the original Tyr and Trp residues. Inturn, Phe in position CD1 displays significant rotamericheterogeneity, including an outward position with respect tothe heme plane. The relatively low resolution did not allow aclear discrimination of the iron bound ligand, although, also inagreement with previous data on Tf-trHb and Bs-trHb, acetatefrom crystallization buffer is most likely the heme ligand. Otherpositive densities are present within the cavity, which are mostlikely water molecules residing on top of the iron bound ligand.The presence of water molecules within the pocket is notsurprising, in spite of the highly hydrophobic character of thedistal pocket in the triple mutant. In fact, previous UV−vis andresonance Raman investigations clearly indicated that the triplemutant, in solution at neutral pH values, displays the typicalspectrum of an aquo-met ferric heme derivative.45 The staticferrous CO adduct displays a single peak at 1955 cm−1, typicalof an unconstrained FeCO moiety and very different fromthe highly constrained, splitted band (1945 and 1920 cm−1)typical of the WT protein. On this basis, we expected that thetriple mutation would have had a very strong effect on thedynamics of recombination of CO, since in the mutant theligand inclusive hydrogen bond network is abolished and theoverall polarity of the distal pocket is significantly altered. Thepresent results, however, indicate that when restricting theanalysis to the short nanosecond time scale, the WT proteinand the triple mutant behave very similarly. In fact, in bothcases, global analysis of the time-resolved spectra recorded inthe infrared region up to two nanosecond pump−probe delayresulted in two kinetic components, on the order of 250 ps and1.4 ns, respectively. Although the kinetic constants obtainedfrom the fitting of the time-resolved data in the two proteins arebasically the same, the relative amount of the picosecondsdynamic phase is different for the mutant protein with respectto the WT one (30% vs 40%). Furthermore, as can be noted byinspecting Figure 7, if the kinetic traces of the mutant and WTproteins are normalized on the long tail, there is a notabledifference in the intensity of the signal at time zero. With thereasonable assumption that the photodissociation yield is thesame in the two proteins, this observation can be rationalizedby assuming that the efficiency of geminate recombination inthe WT is higher than what observed for the mutant. Thiswould imply that a higher amount of CO escapes to the firstcavity in the mutant compared to the WT on a short time scale.However, when considering a longer time interval, spanning upto the millisecond time scale, the total CO recombinationefficiency is higher in the 3F mutant, as reported in a recentstudy.4

In the WT Tf -trHb, the presence of a ligand inclusivehydrogen bond network, in which the three polar amino acidsTrpG8, TyrCD1, and TyrB10 are mostly implicated, representa considerable barrier to ligand escape, which explains theobserved efficient geminate rebinding.1 Furthermore, in WT, ithas been observed that in the “open” conformation, where CO

Figure 9. Free CO normalized spectra (obtained exciting the sampleswith 550 nm laser pulse) taken at 30 ps delay time for Tf-trHb (greenline), 3F-Tf-trHb (black line) and Bs-trHb (red line). ν0 indicates thecentral frequency of the gaseous CO absorption band located at 2143cm−1 (for more details, see Figure S2 in the SI section).

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179214

is stabilized by a single hydrogen bond with TrpG8(characterized by the 1940 cm−1 infrared band) the 250 pskinetic component has a high contribution (about 40%). Thisresult has been rationalized by assuming that the proteinenvironment act as a cage around the photolyzed CO,preventing its escape toward the solvent. In the discussion ofthe WT kinetics, an important role has been recognized to astructural water molecule which is hydrogen bonded toTyrCD1 in the open conformation. Since the weight of thefast picoseconds component resulted lower in case of the“closed” conformation, where the iron coordinate CO isstabilized by two H-bonds donated by TrpG8 and TyrCD1(1920 cm−1 band), it was concluded that there is no directconnection between the number of H-bonds present in thedistal pocket and the caging effect on the photolyzed CO.1 Theresults discussed in this work concerning the 3F mutant areperfectly in line with this conclusion. In fact, in 3F-Tf-trHb, nostabilization of the CO via formation of hydrogen bonds ispossible and most probably the conserved water molecule isdisplaced in this nonpolar distal heme pocket. There is onlyone conformer for which the heme-bound CO stretchingvibration is located at 1955 cm−1. The ultrafast geminaterebinding in this mutant protein is remarkably strong, despitethe absence of the hydrogen-bonding network in the distalpocket, showing that the reactivity of dissociated CO towardthe heme is not strongly changed. Similar findings have beenreported for Mycobacterium tubercolosis trHb and several of itsdistal pocket mutants by Vos and co-workers.19 Also in thatcase, a geminate rebinding on the nanosecond time scale,accounting for 95% of the total recombination, was found bothfor the WT protein and the mutants. The rebinding propertieswere unexpectedly robust against the mutations of the distalpocket residues, and the interaction of the ligand with structuralwater molecules as well as its rotational freedom were indicatedto play a role in the high reactivity of the ligand toward theheme. These results, taken together with those presented in thisstudy show that, at least in the nanosecond time scale, theoccurrence of fast geminate rebinding is regulated principally bya caging effect exerted by the distal pocket residues around thephotolyzed ligand. In the currently analyzed mutant, in fact,bulky polar Trp and Tyr residues have been replaced bynonpolar but equally bulky Phe. This certainly influences thepolarity of the pocket, but probably has a negligible effect onthe effective volume of the cavity where the photodissociatedCO resides soon after being detached by the heme iron. Thisdetermines a very similar geminate recombination in all cases.However, if the photodissociation dynamics is analyzed on alonger time scale, the effect of mutations can becomesubstantial. In the case of Tf -trHb, it has been recentlyshown that when extending the study of the CO recombinationdynamics up to the millisecond time scale, the total amount ofgeminate recombination is substantially higher in the 3Fmutant if compared to the WT protein.4 Molecular dynamicssimulations allowed us to identify three major cavities inside theprotein where the photolyzed CO is temporarily located beforerecombining with the heme or escape toward the solvent, andto set up a kinetic model describing both the migrationbetween the cavities and the geminate and bimolecularrecombination processes. It has been shown that while thedynamics of rebinding inside the first cavity, correspondingwith the distal pocket, are only slightly influenced by themutation of TrpG8, TyrCD1, and TyrB10, the kinetics ofmigration of the free CO among the different cavities are highly

perturbed. In particular, the back kinetic constant for themigration between the second identified cavity and the distalpocket is about 2 orders of magnitude higher in the mutantthan in the WT. This finding implies that in the mutant, thefree CO molecule may gain access and resides longer in the firstcavity, closest to the heme, thus taking into account theobserved higher amount of geminate rebinding on a longermilliseconds time scale.

5. CONCLUSIONS

The inspection of the transient infrared absorption band of thephotolyzed CO molecule has been demonstrated to providekey indications on the degree of homogeneity of the distalpocket and on the confinement exerted on the released ligand.The analysis of the CO recombination dynamics in 3F-Tf -trHbhas shown that the ultrafast geminate rebinding, in this mutantprotein, is remarkably strong despite the absence of thehydrogen-bonding network and that the reactivity ofdissociated CO toward the heme is not strongly changed bythe mutations on the analyzed time scale, up to 2 ns. Thisfinding indicates that on this time scale, the major factorregulating the occurrence of fast geminate rebinding is theeffective volume of the distal pocket and the caging effectexerted by heme surrounding residues on the free CO,irrespective of their polarity or hydrogen bonding properties.The qualitative comparison of the free CO absorption band

of the 3F mutant with those of the WT and another similartruncated hemoglobin, Bs-trHb, shows that the substitution ofthe three polar amino acids TrpG8, TyrB10, and TyrCD1 withPhe residues leads to a nonpolar and more homogeneous distalpocket, which gives rise to a narrower CO band in the mutantprotein respect to the WT. Furthermore, the free COabsorption band of all the three compared proteins is narrowerthan that observed for CO dissolved in organic solvents and theshifts relative to the gaseous CO central frequency are morepronounced for the CO docked in the protein matrix respect tothe CO dissolved in solvents. These observations lead toconclude that the protein matrix and especially the hemepocket flexibility play a very important role in controlling thereactivity of the active site respect to the exogenous ligands. Inconclusion it appears that, although the mutations have littleinfluence on the geminate rebinding within the fast ns timescale, they affect the total amount of geminate recombination,as they modify the communication channels between thedifferent cavities inside the protein, most likely increasing theirtotal volume.

■ ASSOCIATED CONTENT

*S Supporting InformationTime-resolved spectra of 3F-Tf-trHb iron-bound CO; compar-ison between the kinetics of free CO and the time evolution ofthe integrated area of the CO absorption bands in 3F-Tf-trHb;baseline subtraction procedure for data at 2050−2200 cm−1;FT-IR spectra of gaseous CO dissolved in different organicsolvents; crystal parameters, data collection, and refinementstatistics. This material is available free of charge via theInternet at http://pubs.acs.org.

■ AUTHOR INFORMATION

Corresponding Author*E-mail: didonato@lens.unifi.it.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179215

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe acknowledge the Helmholtz-Zentrum BerlinElectronstorage ring BESSY II for provision of synchrotron radiation atbeamline BL 14-1, the italian MIUR (grants FIRBFuturo inRicerca 2010 RBFR109ZHQ and RBFR10Y5VW), RegioneToscana through PORFSE 2007-2013 project EPHODS) andProject EFOR-CABIR.

■ REFERENCES(1) Lapini, A.; Di Donato, M.; Patrizi, B.; Marcelli, A.; Lima, M.;Righini, R.; Foggi, P.; Sciamanna, N.; Boffi, A. Carbon MonoxideRecombination Dynamics in Truncated Hemoglobins Studied withVisible-Pump MidIR-Probe Spectroscopy. J. Phys. Chem. B 2012, 116(30), 8753−8761.(2) Marcelli, A.; Bustamante, J. P.; Feis, A.; Bonamore, A.; Boffi, A.;Gellini, C.; Salvi, P. R.; Estrin, D. A.; Bruno, S.; Viappiani, C.; Foggi, P.Following Ligand Migration Pathways from Picoseconds to Milli-seconds in Type II Truncated Hemoglobin from Thermobif ida fusca.PLoS One 2012, 7 (7), e39884.(3) Feis, A.; Lapini, A.; Catacchio, B.; Brogioni, S.; Foggi, P.;Chiancone, E.; Boffi, A.; Smulevich, G. Unusually Strong H-Bondingto the Heme Ligand and Fast Geminate Recombination Dynamics ofthe Carbon Monoxide Complex of Bacillus subtilis TruncatedHemoglobin. Biochemistry 2007, 47 (3), 902−910.(4) Bustamante, J. P.; Abbruzzetti, S.; Marcelli, A.; Gauto, D.; Boechi,L.; Bonamore, A.; Boffi, A.; Bruno, S.; Feis, A.; Foggi, P.; Estrin, D. A.;Viappiani, C. Ligand Uptake Modulation by Internal Water Moleculesand Hydrophobic Cavities in Hemoglobins. J. Phys.Chem. B 2014, 118(15), 1234−1254.(5) Vos, M. H. Ultrafast Dynamics of Ligands within Heme Proteins.Biochim. Biophys. Acta (BBA)− Bioenergetics 2008, 1777 (1), 15−31.(6) Anfinrud, P. A.; Han, C.; Hochstrasser, R. M. Direct Observationsof Ligand Dynamics in Hemoglobin by Subpicosecond InfraredSpectroscopy. Proc. Natl. Acad. Sci. U. S. A. 1989, 86 (21), 8387.(7) Polack, T.; Ogilvie, J. P.; Franzen, S.; Vos, M. H.; Joffre, M.;Martin, J.-L.; Alexandrou, A. CO Vibration as a Probe of LigandDissociation and Transfer in Myoglobin. Phys. Rev. Lett. 2004, 93 (1),018102−1-018102−4.(8) Austin, R. H.; Beeson, K. W.; Eisenstein, L.; Frauenfelder, H. I.;Gunsalus, C. Dynamics of Ligand Binding to Myoglobin. Biochemistry1975, 14 (24), 5355−5373.(9) Spiro, T. G.; Wasbotten, I. H. CO as a Vibrational Probe ofHeme Protein Active Sites. J. Inorg. Biochem. 2005, 99 (1), 34−44.(10) Touron Touceda, P.; Mosquera Vazquez, S.; Lima, M.; Lapini,A.; Foggi, P.; Dei, A.; Righini, R. Transient Infrared Spectroscopy: ANew Approach to Investigate Valence Tautomerism. Phys. Chem.Chem. Phys. 2012, 14 (2), 1038−1047.(11) Emsley, P.; C, K. Coot: Model-Building Tools for MolecularGraphics. Acta Crystallogr. D Biol. Crystallogr. 2004, No. 60, 2126−2132.(12) Wittenberg, J. B.; Bolognesi, M.; Wittenberg, B. A.; Guertin, M.Truncated Hemoglobins: A New Family of Hemoglobins WidelyDistributed in Bacteria, Unicellular Eukaryotes, and Plants. J. Biol.Chem. 2002, 277 (2), 871−874.(13) Wu, G.; Wainwright, L. M.; Poole, R. K. Microbial Globins. InAdvances in Microbial Physiology; Academic Press: New York, 2003;Vol. 47, pp 255−310.(14) Guallar, V.; Lu, C.; Borrelli, K.; Egawa, T.; Yeh, S.-R. LigandMigration in the Truncated Hemoglobin-II from Mycobacteriumtuberculosis. J. Biol. Chem. 2009, 284 (5), 3106−3116.(15) Ouellet, H.; Milani, M.; LaBarre, M.; Bolognesi, M.; Couture,M.; Guertin, M. The Roles of Tyr(CD1) and Trp(G8) inMycobacterium tuberculosis Truncated Hemoglobin O in LigandBinding and on the Heme Distal Site Architecture. Biochemistry2007, 46 (41), 11440−11450.

(16) Bonamore, A.; Ilari, A.; Giangiacomo, L.; Bellelli, A.; Morea, V.;Boffi, A. A Novel Thermostable Hemoglobin from the Actino-bacterium Thermobif ida fusca. FEBS J. 2005, 272 (16), 4189−4201.(17) Droghetti, E.; Nicoletti, F. P.; Bonamore, A.; Boechi, L.; ArroyoManIf̀ez, P.; Estrin, D. A.; Boffi, A.; Smulevich, G.; Feis, A. HemePocket Structural Properties of a Bacterial Truncated Hemoglobinfrom Thermobif ida fusca. Biochemistry 2010, 49 (49), 10394−10402.(18) Yoshimura, H.; Yoshioka, S.; Kobayashi, K.; Ohta, T.; Uchida,T.; Kubo, M.; Kitagawa, T.; Aono, S. Specific Hydrogen-BondingNetworks Responsible for Selective O2 Sensing of the Oxygen SensorProtein HemAT from Bacillus subtilis. Biochemistry 2006, 45 (27),8301−8307.(19) Jasaitis, A.; Ouellet, H.; Lambry, J.-C.; Martin, J.-L.; Friedman, J.M.; Guertin, M.; Vos, M. H. Ultrafast Heme−Ligand Recombinationin Truncated Hemoglobin HbO from Mycobacterium tuberculosis: ALigand Cage. Chem. Phys. 2012, 396, 10−16.(20) Guallar, V.; Lu, C.; Borrelli, K.; Egawa, T.; Yeh, S. R. LigandMigration in the Truncated Hemoglobin-II from Mycobacteriumtuberculosis: The Role of G8 Tryptophan. J. Biol. Chem. 2009, 284(5), 3106−3116.(21) Schotte, F.; Cho, H. S.; Soman, J.; Wulff, M.; Olson, J. S.;Anfinrud, P. A. Real-time Tracking of CO Migration and Minding inthe α and β Subunits of Human Hemoglobin via 150-ps Time-Resolved Laue Crystallography. Chem. Phys. 2013, 422, 98−106.(22) Pettersen, E. F.; G, T. D.; Huang, C. C.; Couch, G. S.;Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera–aVisualization System for Exploratory Research and Analysis. J. Comput.Chem. 2004, 25 (13), 1605−1612.(23) Otwinowski, Z.; M, W. Processing of X-ray Diffraction DataCollected in Oscillation Mode. Meth. Enzymol. 1997, 276, 307−326.(24) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.;Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A. G.W.; McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N. S.;Potterton, E. A.; Powell, H. R.; Read, R. J.; Vagin, A.; Wilson, K. S.Overview of the CCP4 Suite and Current Developments. ActaCrystallogr. D Biol. Crystallogr. 2011, 67 (4), 235−242.(25) Matthews, B. W. Solvent Content of Protein Crystals. J. Mol.Biol. 1968, 33, 491−497.(26) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement ofMacromolecular Structures by the Maximum-likelihood Method. ActaCrystallogr. D Biol. Crystallogr. 1997, 53, 240−255.(27) Hendler, R. W.; Shrager, R. I. Deconvolutions Based on SingularValue Decomposition and the Pseudoinverse: A Guide for Beginners.J. Biochem. Biophys. Meth. 1994, 28 (1), 1−33.(28) Henry, E. R. The Use of Matrix Methods in the Modeling ofSpectroscopic Data Sets. Biophys. J. 1997, 72 (2), 652−673.(29) Henry, E. R.; Hofrichter, J.; Ludwig Brand, M. L. J. SingularValue Decomposition: Application to Analysis of Experimental Data.In Methods in Enzymology; Academic Press: New York, 1992; Vol. 210,pp 129−192.(30) Dioumaev, A. K. Evaluation of Intrinsic Chemical Kinetics andTransient Product Spectra from Time-resolved Spectroscopic Data.Biophys. Chem. 1997, 67, 1.(31) Bonneau, R.; Wirz, J.; Zuberbuhler, A. D. Methods for theAnalysis of Transient Absorbance Data (Technical Report). Pure Appl.Chem. 1997, 69, 979−992.(32) Henry, E. R. The Use of Matrix Methods in the Modeling ofSpectroscopic Data Sets. Biophys. J. 1997, 72, 652.(33) Franzen, S.; Kiger, L.; Poyart, C.; Martin, J. L. Heme PhotolysisOccurs by Ultrafast Excited State Metal-to-Ring Charge Transfer.Biophys. J.l 2001, 80, 2372.(34) van Stokkum, I. H. M.; Scherer, T.; Brouwer, A. M.; Verhoeven,J. W. Conformational Dynamics of Flexibly and Semirigidly BridgedElectron Donor−Acceptor Systems as Revealed by SpectrotemporalParametrization of Fluorescence. J. Phys. Chem. 1994, 98 (3), 852−866.(35) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Globaland Target Analysis of Time-Resolved Spectra. Biochim. Biophys. Acta(BBA)-Bioenergetics 2004, 1657, 82.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179216

(36) Satzger, H.; Zinth, W. Visualization of Transient AbsorptionDynamicsTowards a Qualitative View of Complex ReactionKinetics. Chem. Phys. 2003, 295 (3), 287−295.(37) Neumann, K.; Verhoefen, M.-K.; Weber, I.; Glaubitz, C.;Wachtveitl, J. Initial Reaction Dynamics of Proteorhodopsin Observedby Femtosecond Infrared and Visible Spectroscopy. Biophys. J. 2008,94 (12), 4796−4807.(38) Mullen, K. M.; van Stokkum, I. H. M. An Introduction to the“Special Volume Spectroscopy and Chemometrics in R”. J. Stat. Soft.2007, 18 (1), 1−5.(39) Mullen, K. M.; van Stokkum, I. H. M. TIMP: An R Package forModeling Multi-way Spectroscopic Measurements. J. Stat. Soft. 2007,18 (3), 1−46.(40) Snellenburg, J.; Laptenok, S. P.; Seger, R.; Mullen, K. M.; vanStokkum, I. H. M. Glotaran: A Java Based Graphical User Interface forthe R Package TIMP. J. Stat. Soft. 2012, 49 (3), 1−22.(41) Snellenburg, J. VU University: Amsterdam, http://glotaran.org/.(42) Lim, M.; Jackson, T. A.; Anfinrud, P. A. Ultrafast Rotation andTrapping of Carbon Monoxide Dissociated from Myoglobin. Nat.Struct. Biol. 1997, 4, 209−214.(43) Lim, M.; Jackson, T. A.; Anfinrud, P. A. Vibrational Spectrum ofCO after Photodissociation from Heme: Evidence for a LigandDocking Site in the Heme Pocket of Hemoglobin and Myoglobin. J.Chem. Phys. 1995, 102, 4355−4366.(44) Alben, J. O.; Beece, D.; Bowne, S. F.; Doster, W.; Eisenstein, L.;Frauenfelder, H.; Good, D.; McDonald, J. D.; Marden, M. C.; Moh, P.P.; et al. Infrared Spectroscopy of Photodissociated Carboxymyoglobinat Low Temperatures. Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 3744−3748.(45) Nicoletti, F. P.; Droghetti, E.; Howes, B. D.; Bustamante, J. P.;Bonamore, A.; Sciamanna, N.; Estrin, D. A.; Feis, A.; Boffi, A.;Smulevich, G. H-Bonding Networks of the Distal residues and WaterMolecules in the Active Site of Thermobif ida fusca Hemoglobin.Biochim. Biophys. Acta−Proteins Proteomics 2013, 1834, 1901−1909.

The Journal of Physical Chemistry B Article

dx.doi.org/10.1021/jp504499b | J. Phys. Chem. B 2014, 118, 9209−92179217