1

Structural Dynamics in the Active Site of Murine Neuroglobin

and its Effects on Ligand Binding

Karin Nienhaus‡, Jan M. Kriegl‡, and G. Ulrich Nienhaus ‡†*

March 11, 2004

‡Department of Biophysics

University of Ulm

Albert-Einstein-Allee 11

89081 Ulm, Germany

§Department of Physics

University of Illinois at Urbana-Champaign

1110 W. Green

Urbana, Il 61801, USA

Running Title: Structural Dynamics and Ligand Binding to Ngb

* To whom correspondence should be addressed: E-mail: uli@uiuc.edu

† This work was supported by grants from the Deutsche Forschungsgemeinschaft (Ni-291/3)

and the ‘Fonds der Chemischen Industrie’.

Abbreviations: Ngb, neuroglobin; Mb, myoglobin.

JBC Papers in Press. Published on March 11, 2004 as Manuscript M401561200

Copyright 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

2

We have examined the effects of active site residues on ligand binding to the heme iron

of mouse neuroglobin using steady-state and time-resolved visible spectroscopy.

Absorption spectra of the native protein, mutants H64L and K67L and double mutant

H64L-K67L were recorded for the ferric and ferrous states over a wide pH range (4 −

11), which allowed us to identify a number of different species with different ligands at

the sixth coordination, to characterize their spectroscopic properties and to determine

the pK values of active site residues. In flash photolysis experiments on CO-ligated

samples, reaction intermediates and the competition of ligands for the sixth coordination

were studied. These data provide insights into structural changes in the active site and

the role of the key residues His64 and Lys67. His64 interferes with exogenous ligand

access to the heme iron. Lys67 sequesters the distal pocket from the solvent. The heme

iron is very reactive, as inferred from the fast ligand binding kinetics and the ability to

bind water or hydroxyl ligands to the ferrous heme. Fast bond formation favors

geminate rebinding, yet, the large fraction of bimolecular rebinding observed in the

kinetics implies that ligand escape from the distal pocket is highly efficient. Even slight

pH variations cause pronounced changes in the association rate of exogenous ligands

near physiological pH, which may be important in functional processes.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

3

Globins are small respiratory proteins that reversibly bind dioxygen and other small ligands at

the central iron of a heme prosthetic group that is embedded in a highly conserved α-helical

globin fold. They are widely distributed over many taxa, including unicellular eukaryotes,

plants, fungi, and animals (1). Hemoglobin (Hb) and myoglobin (Mb) are the most prominent

members of this protein family (2).

Recently, two new members have been discovered in the vertebrate globin family,

cytoglobin, also known as histoglobin, and neuroglobin (Ngb) (3-6). Whereas cytoglobin is

ubiquitously found in vertebrate tissue, Ngb is predominantly expressed in neuronal cells of

the brain (4,5,7,8). Phylogenetic analysis shows that Ngb is an ancient protein that existed

long before the genes encoding Mb and Hb diverged. Nevertheless, it displays all key

determinants of the globin fold (9). Ngb consists of a single chain of 151 amino acids and

shares only 20 – 25% sequence identity with vertebrate Mb or Hb. Human and murine Ngb

have 94% identical residues and are thus more similar than Hb and Mb in these species (77 –

85%) (5).

In the absence of an exogenous ligand, the heme iron is hexacoordinate in both ferric

(Fe3+, met) and ferrous (Fe2+, deoxy) Ngb (5,10-12), with His64 and His96 as axial ligands, as

has been confirmed by recent x-ray structure analyses of human (13) and murine (14) met

Ngb. Similar hexacoordinate globins have also been isolated from bacteria and unicellular

eukaryotes (truncated hemoglobins, trHbs (15,16)) as well as from plants (non-symbiotic

hemoglobins, nsHbs (17-19)) and insects (20). The exogenous ligand replaces the endogenous

one upon binding at the sixth coordination, and therefore, hexacoordination has been

considered as a novel mechanism for regulating ligand binding affinity to the heme (13). The

physiological role of hexacoordinate globins has not yet been clarified. For Ngb, the low

expression level and moderate oxygen affinity (half saturation pressure P50 ≈ 2.0 torr (5))

suggests a function different from simple oxygen storage and transport (21). Recent studies

have reported an up-regulation of Ngb under conditions of hypoxia or ischemia in vitro

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4

(21,22) and in vivo (23), thus promoting neuronal survival. The induction of Ngb expression

by hemin indicates that more than one signal transduction pathway is involved in the

regulation of Ngb expression, because heme availability and hypoxia-induced pathways are

controlled through different mechanisms (24). Moreover, ferric (but not O2 ligated ferrous)

human Ngb interacts with the GDP-bound form of the α subunit of the heterotrimeric G

protein, thereby preventing neuronal death (25). All these results suggest that Ngb is a sensor

responsive to oxidative stress.

To understand the function of Ngb in the vertebrate brain and that of hexacoordinate

globins in general, a thorough investigation of the structure, dynamics and ligand binding

(equilibrium and kinetic) properties is a prerequisite. Figure 1 shows the essential features of

the active site of murine met Ngb, as determined at 105 K (14). The heme is inserted into the

globin in two different orientations, as already suggested from NMR investigations by La Mar

et al. (26), which also affects the axial residues His64 and His96. Structural heterogeneity in

the distal heme pocket was also apparent from infrared spectroscopy on CO-ligated Ngb (27).

Kinetic studies have revealed extremely fast CO and O2 association to pentacoordinate

deoxy Ngb and extremely slow binding to the hexacoordinate deoxy form. The latter process

is governed by thermal dissociation of the His64 sidechain from the sixth coordination

(10,12,27). Additional processes involving other amino acid sidechains and/or ligands were

inferred from infrared and optical absorption (27) and Raman (28) measurements. To examine

structural and dynamic aspects of ligand binding, we have prepared Ngb mutants H64L, K67L

and H64L-K67L. Here we present steady-state optical absorption spectroscopy on the deoxy,

met and CO-ligated forms and time-resolved spectroscopy on the CO-ligated proteins.

Measurements over a wide range of pH enabled us to examine the competition of endogenous

and exogenous ligands for the sixth heme coordination as a function of the protonation states

of the residues in the vicinity of the active site.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

5

Materials and Methods

Site-directed Mutagenesis. Murine Ngb (wt) was produced in Escherichia coli strain BL21

(DE3)pLys by using the pET3a expression plasmid (27). Mutants H64L and K67L and double

mutant H64L-K67L were constructed using the Quikchange mutagenesis kit (Stratagene

Europe, Amsterdam) following the manufacturer’s protocol. Custom designed primers were

purchased from MWG (MWG-Biotech GmbH, Ebersberg, Germany). Mutants were

expressed and purified as described for the wt protein (27).

Sample preparation. Optical absorption spectra and flash photolysis kinetics at ambient

temperature were measured on dilute aqueous samples, prepared by dissolving lyophilized

protein in 100 mM sodium phosphate/citrate (pH 4 − 6), sodium phosphate (pH 6.2 − 8.6) and

sodium carbonate buffer (pH > 8.6) to a final concentration of ~10 µM. For cryospectroscopy,

the protein was dissolved in 75%glycerol/25% potassium phosphate buffer (vol/vol). To

produce NgbCO samples, the solutions were equilibrated with 0.05 or 1 atm CO partial

pressure under anaerobic conditions, and a tenfold excess of sodium dithionite was

subsequently added. Six-coordinate deoxy Ngb samples were prepared by adding dithionite

solution anaerobically to ferric Ngb solutions after thorough purging with N2 gas.

Absorption spectroscopy. Spectra at ambient temperature were taken with a Cary 1 E

spectrometer (Varian, Darmstadt, Germany) at a resolution of 1 nm. Low-temperature spectra

were recorded on an OLIS-modified Cary 14 spectrometer (On-Line Instrument Systems,

Bogart, GA) at a resolution of 0.3 nm. Samples were cooled with a closed-cycle helium

cryostat (model 22, CTI Cryogenics, Mansfield, MA) equipped with a Lakeshore Cryotronics

model 330 digital temperature controller. To observe photolysis-induced spectral changes,

samples were illuminated at low temperature for 10 min prior to data collection by using a

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6

frequency-doubled Nd:YAG laser emitting ~300 mW at 532 nm (model Forte 530-300, Laser

Quantum, Manchester, UK).

Flash Photolysis Experiments. In our home-built flash photolysis apparatus, photolysis was

achieved by a saturating 6-ns (full width at half maximum) pulse from a frequency doubled,

Q-switched Nd:YAG laser (model Surelite II, Continuum, Santa Clara, CA). Light-induced

absorbance changes were monitored in the Soret region using light from a tungsten source

(model A 1010, PTI, Brunswick, NJ). The wavelength of the monitoring beam was adjusted

by a monochromator. The transmitted light intensity was measured with a photomultiplier

tube (model R5600U, Hamamatsu Corp., Middlesex, NJ) and recorded with a digital storage

oscilloscope from 10 ns to 50 µs (model TDS 520, Tektronix, Wilsonville, OR) and a home-

built logarithmic time-base digitizer (Wondertoy II) from 2 µs to 100 s. Between 100 and 500

individual transients were averaged for each data set. For single-wavelength kinetics, the

absorbance change was measured at 436 nm and normalized to 1 at the earliest times (30 ns).

We denote the normalized absorbance change by N(t); it is the fraction of molecules that have

not yet rebound a ligand at time t after the photolyzing flash.

Experimental Results

Optical absorption at ambient temperature. The spectra of Ngb mutant K67L in the different

oxidation and ligation states of the heme iron are similar to those of wt Ngb (10). The two (α

and β) bands between 500 and 600 nm are indicative of hexacoordination of the heme iron in

the oxidized (met) and reduced (deoxy) states without exogenous ligand as well as the CO-

ligated state. The peak positions of the various bands are compiled in Table 1. For the CO-

ligated form, a small but significant spectral change with pH was noticed. With decreasing

pH, the Soret (γ) band of wt NgbCO (K67L NgbCO) red-shifts from 415.6 – 416.6 (416.5 –

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

7

417.2) nm according to the Henderson-Hasselbalch relation (Fig. 2 a). This relation gives the

fractional populations of the protonated (acid) and deprotonated (base) species, here denoted

by c+ and c0, as a function of pH,

pK)(pH0 1011)pH(1)pH( −+ +

=−= ncc . (1)

For the Henderson-Hasselbalch relation, which describes a single protonation, the parameter n

equals 1; n > 1 signals a cooperative transition involving more than one protonation. The data

in Fig. 2 a show that a protonating group with a pK of 4.4 ± 0.1 (5.6 ± 0.1) resides in the

vicinity of the heme. For the CO-ligated mutants with leucine at position 64, a similar shift of

the Soret band was not observed.

In the absence of an exogenous ligand, replacing His64 by leucine results in substantial

pH-dependent spectral changes in both the ferric and ferrous forms (Fig. 3 a, b). At low pH,

the deoxy form of H64L Ngb is characterized by a broad Soret band, with a maximum at 422

nm and a pronounced shoulder at 436 nm (Fig. 3 b). The second-derivative spectrum in Fig. 3

c displays two minima, clearly indicating that the Soret consists of two peaks. The spectral

region between 500 and 600 nm exhibits a broad absorption feature around 550 nm. It

contains multiple bands, as seen from the (most pronounced) minima in the second-derivative

spectra at 530, 560 and 571 nm (Fig. 3 c). The spectra are pH-independent between pH 5 and

8. At more alkaline pH, they change dramatically (Fig. 3 b, c). The Soret band becomes much

larger and narrower; the shoulder at 436 nm disappears completely. Moreover, two narrow

and tall bands at 527 and 557 nm arise from the broad 550 nm absorption. These changes

clearly signal the appearance of a new low-spin heme iron species. In Fig. 2 b, the absorbance

change at 557 nm is plotted as a function of pH after rescaling from 0 – 1. The line through

the data is the result of a non-linear least-squares fit of the data with a Henderson-Hasselbalch

relation, yielding a pK of 9.9 ± 0.1. This behavior suggests that an amino acid near the active

site deprotonates and binds to the heme iron.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

8

At acidic pH, ferric Ngb H64L has a Soret peak at 403 nm (Fig. 3 a). Above 500 nm, the

spectrum exhibits the typical, multi-humped appearance of an aquomet globin species (2),

with the charge-transfer band at 631 nm that is characteristic of a high-spin ferric heme iron

(29). With increasing pH, the Soret shifts to the red by ~6 nm, and concomitantly, a double-

humped spectrum appears above 500 nm, indicating that H2O at the sixth coordination is

replaced by another ligand. Fig. 2 b shows the scaled absorbance change at 409 nm in the

Soret as a function of pH. The pH dependence is in excellent agreement with a Henderson-

Hasselbalch relation with pK = 7.3 ± 0.1. Above pH 9, the absorption at 409 nm decreases

slightly, which signals a protonation with pK ≈ 10. This is likely the same protonation that has

already been much more obvious in the deoxy spectra.

The ferric double mutant H64L-K67L also shows the acidic/alkaline transition (Fig. 3 d).

From the change in absorbance at 403 nm, a pK of 8.8 ± 0.3 was determined (Fig. 2 c). At pH

> 9, a second transition manifests itself in a pronounced and rather abrupt shift of the Soret

band from 407 – 421 nm (shown scaled from 0 – 1 in Fig. 2 c). The fit with Eq. 1 yielded a

cooperativity parameter n = 3.5. The α and β bands change very little, however (Fig. 3 d). By

contrast, the spectra of the ferrous species appear essentially pH-independent (Fig. 3 e). The

second derivative spectrum (Fig. 3 f) only shows slight shifts of the bands with pH.

Optical spectra at cryogenic temperature. After photodissociation of CO or other small,

gaseous ligands at ambient temperature, a pentacoordinate deoxy Ngb forms, which is

converted within milliseconds to the hexacoordinate deoxy species by bond formation

between the heme iron and the His64 imidazole (10,27). At cryogenic temperature, however,

large-scale protein motions are frozen in (30,31). Therefore, the His64 imidazole side chain is

immobilized, and the CO rebinds to the pentacoordinate heme iron from the distal heme

pocket. We have shown earlier that the enthalpy barrier for CO binding to pentacoordinate

Ngb is extremely low in comparison with other heme proteins (27,32). As a consequence,

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

9

only at very low temperatures can a significant amount of pentacoordinate NgbCO

photoproduct be trapped for longer periods of time after photolysis. Figure 4 a shows the

Soret region of the NgbCO absorption spectrum before and after illumination at 10 K. Upon

photodissociation, the intensity of the Soret peak at 414 nm decreases, concomitant with the

appearance of the broad absorption of the pentacoordinate species with a maximum at 438

nm, as is evident from the difference spectrum. Essentially identical spectral changes are seen

for mutant H64L (Fig. 4 b), consistent with our expectation that His64 cannot bind to the sixth

coordination in the native protein after photodissociation of CO at low temperatures.

To further elucidate the complicated spectra of ferrous H64L Ngb at room temperature

(Fig. 3 b), we also performed cryospectroscopy on this species. The spectra in Fig. 4 c were

collected on a sample at pH 7.5. They show a pronounced and completely reversible

temperature effect. With decreasing temperature, the peak at 438 nm gets weaker at the

expense of the band at 418 nm. After laser illumination, this change is essentially reverted,

which indicates that the band at 418 nm belongs to a hexacoordinate species with a small,

photolyzable ligand at the sixth coordination, whereas the band at 438 nm belongs to

pentacoordinate, high-spin heme. Comparison of samples in the pH range 6 – 10 showed that

the hexacoordinate species grows at the expense of the pentacoordinate form towards higher

pH (data not shown). A third, rather weak feature at 428 nm is only visible in the better-

resolved spectra at 10 K, both before and after photolysis (Fig. 4 c, d). However, it is too

weak to decide unambiguously from the data if its fraction decreases or increases upon

photolysis.

CO binding after photodissociation at ambient temperature. The ligand binding reaction in

NgbCO after photodissociation is a complicated process that can best be followed from the

time dependence of entire optical spectra. Figure 5 shows a contour plot of the spectral

changes that occur after photodissociation of wt NgbCO, pH 4, 0.05 bar CO, logarithmically

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

10

plotted as a function of time between 30 ns and 10 s. Negative contours (dashed lines) on the

left represent changes in the Soret band of NgbCO and positive contours (solid lines) on the

right changes in the Soret band of the species that occur upon CO photodissociation. A

geminate phase due to CO molecules that do not exit into the solvent after photodissociation

has been observed at 275 K at times below 100 ns (27). At room temperature, this process is

faster than 30 ns and thus outside of the time window of our instrument. A large fraction

rebinds in a bimolecular process on the time scale of ~10−3 s, as is apparent from the decrease

in the difference spectrum at that time. This process does not complete, however, but stalls for

many orders of magnitude in time because the endogenous His64 binds to the heme iron, thus

blocking the active site. Only after thermal dissociation of His64, which occurs on the 1-s

time scale at pH 4, can the rest of the CO finally bind and the difference spectrum decay to

zero. This interpretation is evident from the spectral changes in the Soret band of the

photoproducts. The positive signal at early times is centered on ~432 nm but shifts to 425 nm

due to the binding of His64 to the sixth coordination.

In most experiments, we take kinetic data only at a fixed wavelength, normally near the

maximum of the pentacoordinate species at 436 nm (solid line in Fig. 5). In Fig. 6, we

compare CO rebinding in wt Ngb and its mutants with wt Mb and Mb H64L. All kinetic

traces were taken on samples in 0.1 M buffer, pH 8, equilibrated with CO at a partial pressure

of 0.05 bar, except for a single trace with wt NgbCO (open circles, 1 bar CO). In the time

window of our experiments, the Ngb kinetics at 436 nm show two decays, characterized by

apparent rate coefficients λ1 and λ2 and amplitudes N1 and N2. The first step in the kinetics at

1 bar is much faster than at 0.05 bar, as expected for bimolecular recombination of CO to the

pentacoordinate deoxy form. The lines represent double-exponential fits to the data, the

resulting second-order rate coefficients for CO binding to the pentacoordinate deoxy species

are summarized in Table 2. Note that the amplitudes in these traces do not represent the

relative amounts of the two transient deoxy species because the first decaying species is

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

11

measured near its absorption maximum whereas the second, hexacoordinate species is

measured in the wing of its absorption band (Fig. 5).

CO binding in mutant K67L is similar to wt Ngb, slightly faster but also biphasic, with

two well-separated steps. By contrast, mutants H64L and H64L-K67L yield essentially

monophasic kinetic traces. Their rate coefficients are ~4 – 5-fold larger than the one of the

wild-type protein (Table 2). Consistent with the kinetic behavior at low temperature (27), the

association rate for CO binding to wild-type Ngb at ambient temperature is ~100 × higher

compared to wild type sperm whale Mb and even 2 – 3 × higher than in Mb H64L, which

binds ligands the fastest among all Mb mutants studied so far (33).

pH dependence of CO binding. The titration curves in Fig. 2 and the absorption spectra in Fig.

3 reveal a strong pH dependence of the ligation states in Ngb and its mutants. Therefore, we

have also measured the absorbance changes of wt NgbCO and mutants at 436 nm after

photodissociation at ambient temperature as a function of pH. Fig. 7 shows kinetic traces of

wt NgbCO for two different CO concentrations in the solvent, obtained by equilibration with

1 bar (Fig. 7 a) and 0.05 bar CO (Fig. 7 b). The larger amplitude of the second step at 0.05 bar

CO is evidence of the competition between CO and His64 for the sixth coordination. The

slower bimolecular CO recombination at 0.05 bar gives the His64 imidazole sidechain a

greater probability of binding at the heme iron. The pH dependence of the kinetics is

complicated. The bimolecular process slows with decreasing pH, the yield of the

hexacoordinate deoxy species varies with pH in a complicated fashion, and the termination of

the second step shifts to shorter times for acidic pH (~4), which reflects destabilization of the

bond between the heme iron and the endogenous His64 ligand. The kinetic traces of mutant

K67L are similar to those of wt Ngb. Bimolecular recombination occurs on similar time scales

and also slows towards low pH, and amplitude and termination of the second step are also pH-

dependent.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

12

NgbCO mutants H64L and H64L-K67L show monophasic recombination kinetics at 1 bar

CO. Only if the concentration of CO above the sample is reduced to 0.05 bar or below, a

second process becomes apparent for mutant H64L. Fig. 7 c shows a few select traces. The

kinetics are pH-independent below pH 9. At pH > 9, CO association to the pentacoordinate

species becomes faster with increasing pH, and a second, slower process occurs. In the double

mutant H64L-K67L, we were unable to detect any pH-dependent changes in the kinetics

between pH 5 and 11.

Fig. 8 shows apparent rate coefficients, λ1, for the first step in the kinetics of wt Ngb and

mutants K67L, H64L, and H64L-K67L between pH 4 and 11. At 1 bar of CO, endogenous

ligand binding is small, and thus λ1 is essentially identical with the CO association rate

coefficient. For wt NgbCO, there is a pronounced increase of the CO binding rate by a factor

of ~4 towards higher pH values, well described by a Henderson-Hasselbalch relation with pK

6.0 ± 0.1. For mutant K67L, a similar behavior is observed, with slightly larger amplitude of

the transition and a shift to pK 6.3 ± 0.1. By contrast, a systematic variation of the rate with

pH could not be detected in this pH range for mutants H64L and H64L-K67L. To observe the

Henderson-Hasselbalch dependence for the overall rate, it is obvious that the protonated and

unprotonated species must have different binding rates. Moreover, conformational

fluctuations between them have to be faster than the reaction rate; otherwise, we would

observe two separate kinetic species. The overall rate coefficient λ(pH) is given by the linear

combination,

00 )pH()pH()pH( λλλ cc += ++ . (2)

The data in Fig. 8 yield excellent agreement with this relation. The missing pH dependence of

the mutants lacking His64 suggests that this residue is responsible for the changes.

Figure 9 shows a complicated pH dependence of the amplitude of the second step, N2,

which is proportional to the yield of hexacoordinate Ngb after photodissociation, for wt

NgbCO and mutant K67LCO at 0.05 and 1 bar CO. In addition to the overall increase of the

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

13

yield of hexacoordinate deoxy Ngb with decreasing CO concentration mentioned earlier, there

is up to 10-fold yield variation with pH. The increase in N2 of wt Ngb in the alkaline region

(pH ≈ 10) is absent for the K67L mutant. In the acidic region, the hexacoordinate yield is low

at pH 4, increases to a maximum at pH 6 for wt NgbCO, and afterwards decreases again

towards pH 8. The peak is shifted to slightly higher pH for the K67L mutant. Apparently, the

maximum yield of hexacoordinate species coincides with the pK of the bimolecular CO

association to the pentacoordinate species (Fig. 8).

Discussion

Ngb is a vertebrate globin with the canonical 3-over-3 globin fold (13,14). Besides

hexacoordination of the ferric and ferrous unligated species, it exhibits a few other remarkable

structural features. A huge internal cavity with a volume of ~280 Å3 connects to a large tunnel

through the matrix that may allow for easy ligand migration. Preliminary data indicate that the

heme group shifts deeper into the heme crevice upon exogenous ligand binding. Whereas the

structural data provide only a static picture, the results presented here provide some insights

into the complex, time-dependent interactions among distal site residues His64, Lys67, the

heme iron and the exogenous ligand.

Equilibrium conformations in the presence of His64. The absorption spectra of ferric and

ferrous murine Ngb and mutant K67L show hexacoordination over the entire pH range, with

His64 and His96 as axial ligands (13,14). In the CO-ligated state, the His64 imidazole is

detached from the heme iron, but still located within the distal heme pocket at physiological

pH, as can be inferred from the FTIR spectra of the CO stretching vibration (27). The pH-

dependent Soret shifts with pK 4.4 and 5.6 in wt NgbCO and K67L NgbCO, respectively

(Fig. 2 a), are associated with His64 protonation because they are absent in the mutants

lacking this residue. Further support for this assignment is provided by detailed studies of

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

14

MbCO and its distal pocket mutants, which have proven that His64 protonation occurs in

MbCO with pK 4.5 (34,35). Yang and Phillips (36) have shown that, upon protonation, the

His64 imidazole swings out of the hydrophobic distal heme pocket into the solvent to better

solvate the charge. In NgbCO, the change in population of the CO stretching bands at 1940

and 1970 cm−1 in the FTIR spectra with pK ≈ 4.4 suggest a similar scenario (27). These bands

correspond, respectively, to the A1 and A0 bands in MbCO, which have been associated with

the neutral and charged states of the His64 imidazole (34).

In ferric Ngb, H64 binds to the heme iron with its Nε2 atom. Bond formation impedes

protonation and, consequently, a pH-dependent change in the absorbance spectra with pK ≈

4.4 is absent. From acid denaturation studies of Mb it is known that the covalent bond

between the proximal histidine and the heme iron breaks at pH < 3.5, with subsequent

protonation of the histidine (37,38). A similar scenario is likely for Ngb involving both axial

histidines.

The difference in the pK of His64 between wt and K67L NgbCO indicates destabilization

of the protonated form by ~6.7 kJ/mol in the presence of the charged lysine residue. In water

at 298 K, the imidazole sidechain of histidine has a pK of 6.14 (39,40), with a preference of

4:1 for the proton at Nε2 instead of Nδ1. In a protein, pKs of protonating groups can vary

drastically due to steric and charge interactions within their specific microenvironments. The

pK of His64 in wt NgbCO is rather low for two reasons, (i) electrostatic interaction of His64

with the positive charge of Lys67 and (ii) low accessibility of the apolar protein interior for

solvent. Lys67 forms a salt bridge to a heme propionate group, so that the Lys67 side chain

acts as a stable barrier separating the distal pocket from the solvent. In the K67L mutant, the

neutral leucine side chain cannot form a salt bridge to the heme propionates (14). Therefore,

this modification is expected to facilitate access of solvent molecules to the distal heme

pocket. The pK shift to 5.6 in this mutant, which is closer to the canonical pK of histidine,

indeed suggests a more solvent-accessible environment of His64 in K67L NgbCO. Similar pK

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

15

shifts of His64 have been observed in MbCO upon replacement of the neighboring residue

Arg45 (33).

Equilibrium conformations in the absence of His64. In ferric wt Mb, the distal histidine

stabilizes a water molecule at the sixth coordination (41). In ferric H64L Ngb, a sixth ligand is

obvious from the spectra in Fig. 3 a. The pH-dependent spectral changes reveal a protonation

with pK 7.3 (Fig. 2 b), suggesting heme ligation by H2O at acidic and OH− at alkaline pH.

This acidic-alkaline transition, which has also been observed in human Ngb mutant H64V

(28), is well known from many other ferric heme proteins (2), with pKs varying between 7.4

and 10.9. The pK in Chironomus insect Hbs is similar to the one observed in Ngb. Both

globins also have a positively charged amino acid (arginine instead of lysine) at position 67.

Remarkably, as in murine Ngb, the heme binds to the globin in two orientations in

Chironomus Hb (42,43).

In the absence of His64 as the sixth ligand, one would normally expect a pentacoordinate

high-spin iron in the ferrous deoxy form, with a red-shifted Soret band peaking at 430 – 440

nm. Therefore, we assign the shoulder at 436 nm in the spectra in Fig. 3 b to the five-

coordinate species. The predominant species, however, peaks at 422 nm, which is

characteristic of hexacoordinate low-spin heme. Previously, we had shown that a deoxy Mb

species, with a photodissociable water (or OH–) ligand at the sixth coordination, can be

generated by photoreduction of aquomet Mb at cryogenic temperature (44). In Ngb, the fast

geminate rebinding of CO to the pentacoordinate species at cryogenic temperature (27) and

the fast bimolecular CO binding at ambient temperature (Fig. 6) both indicate that the heme

iron is extremely reactive towards ligands, as is also expected from the iron out-of-plane

displacement towards the distal side by 0.2 Å (14). Therefore, it appears reasonable that Ngb

H64L has already substantial affinity for water/OH− even at room temperature. Further

support comes from the cryospectroscopy data. Upon cooling, the equilibrium shifts from the

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

16

pentacoordinate to the hexacoordinate form (Fig. 4 c), as expected for a bimolecular reaction

for entropic reasons. Moreover, the photoinduced spectral changes further confirm the

presence of a small ligand at the sixth coordination. We have noticed that the band at 418 nm

(in cryosolvent) grows at the expense of the red shoulder as the pH is raised from 6 – 10 (data

not shown), which may suggest that the photolyzable ligand is an OH− ion. The origin of the

third peak at 428 nm is not yet clear; it could be due to binding of water as in photoreduced

met Mb (44).

In both ferric and deoxy Ngb H64L, there is another change in ligation state at alkaline pH

(Fig. 3 b, c). Above pH 9, the spectra strongly resemble those of native deoxy Ngb, with sharp

α and β bands typical of two nitrogenous ligands coordinated axially to the heme iron. We

associate these changes with ligation of the Lys67 amino group to the heme iron after

deprotonation, which happens with a pK of 9.9 (Fig. 2 b). This process was also seen in

Raman spectra of human Ngb by Uno et al. (28). Consistent with this assignment, drastic

changes at high pH are absent in the spectra of ferrous H64L-K67L Ngb (Fig. 3 e, f).

Coordination of the heme iron by non-histidine residues in the distal pocket has also been

reported for Synechocystis hemoglobin (15,45). Anyhow, the distal histidine is clearly the

primary endogenous ligand in Ngb and a variety of other hemoglobins (46-48).

In ferric H64L-K67L Ngb, the alkaline transition occurs with a pK of 8.8, 1.5 pH units

higher than in H64L Ngb. The positively charged Lys67 side chain may assist in binding the

OH− ion; therefore, the pK shift upon replacement of Lys67 by leucine occurs likely because

of the lacking electrostatic stabilization by a positively charged group.

Much to our surprise, another protonation transition was found at even higher pH in ferric

H64L-K67L from the analysis of the Soret shift. This rather sharp transition (Fig. 2 c) cannot

be described by a simple Henderson-Hasselbalch relation involving a single proton (n = 1 in

Eq. 1); it represents a cooperative transition involving 3 – 4 protonating groups. We also

noticed that, at pH > 8.5, the absorption spectra changed very slowly with time and took hours

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

17

to reach equilibrium. These features are characteristic of partial denaturation and subsequent

refolding into a non-native conformation. A similar phenomenon was also observed with the

double mutant H46L-Q43L of Synechocystis hemoglobin (45). There, Hvitved et al.

suggested that the double mutation created a highly unstable distal pocket so that other

residues can bind to the sixth coordination of the heme iron. A likely ligand for

hexacoordination in Ngb H64L-K67L at pH > 8.5 is Arg66 (see Fig. 1).

Ligand binding and conformational changes. Flash photolysis kinetics yield information on

dynamic processes after photodissociation and intermediate states. At ambient temperature,

one usually observes a fast geminate phase on nanosecond to microsecond time scales (in

aqueous solvent). For example, geminate rebinding in MbCO and MbO2 extends out to almost

1 µs, indicating that ligands spend up to 1 µs within the protein after bond cleavage, during

which they either recombine or escape into the solvent (49). In NgbCO, a pronounced

geminate phase extending out to 100 ns was visible at 275 K (27). At higher temperatures,

this process becomes faster than the ~30 ns time resolution of our instrument (Figs. 5, 6). The

fast geminate rebinding reflects the low enthalpy barriers at the heme iron (27). The fact that

solvent rebinding is nevertheless the dominant kinetic process near physiological temperature

indicates that escape of CO from the protein must be very efficient.

In samples equilibrated with 0.05 bar CO, there is essentially no ligand binding between

30 ns and 100 µs (Figs. 5 – 7). The deligated species is characterized by a broad spectrum

around 432 nm. By contrast, the five-coordinate species in NgbCO H64L and H64L-K67L are

both centered on 436 nm. This considerable shift to the blue in the wt NgbCO photoproduct

may indicate partial occupation of the sixth coordination by water after CO escape.

The first step in the kinetics with apparent rate coefficient λ1 arises from the binding of the

exogenous ligand CO from the solvent with rate coefficient λCO, plus binding of the

endogenous ligand (predominantly His64 at physiological pH) with rate coefficient λHis.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

18

Under our experimental conditions, with [CO] >> [Ngb], λCO is a pseudo-first order rate

coefficient proportional to [CO]. In a simple kinetic scheme, in which CO and His64 compete

for the sixth coordination with unique rate coefficients, the apparent rate coefficient, λ1, and

the amplitude of the second step, N2, which is proportional to the yield of His64 binding, are

given by

COHis λλλ +=1 , (3)

COHis

HisconstNλλ

λ+

= .2 . (4)

At 1 bar CO, N2 is small because λCO >> λHis, and λ1 ≈ λCO. CO association is more than two

orders of magnitude faster than in MbCO ((49), Table 1), which again reflects the high

reactivity of the heme iron in Ngb. At 0.05 bar CO, λ1 is much smaller because of the

decreased [CO], proving that CO binding is indeed a bimolecular reaction. Moreover, both

λHis and λCO are of the same order of magnitude, and there is a markedly increased fraction of

endogenous ligands, N2 (Figs. 6 and 7). The second plateau in the kinetics persists up to the

lifetime of the His64-Fe bond. After thermal dissociation, the H64 ligand is finally replaced

by the more strongly bound CO, λ2 ≈ koff (His64). The shortening of the plateau at pH 4 (Fig.

7 a, b) hence signals destabilization of the His64-Fe bond under acidic conditions.

The pH dependence of λ1 (≈ λCO at 1 bar CO) is most interesting. Fig. 8 clearly shows that

λCO increases by more than 4-fold with pH according to Henderson-Hasselbalch relations with

pKs of 6.0 and 6.3 for wt and K67L Ngb, respectively. The comparison with the H64L

mutants in Fig. 8 shows that λCO is much faster and pH-independent if the bulky His64 is

replaced by the apolar leucine. Clearly, His64 must be responsible for the pH dependence of

λCO, and it hinders ligand access to the binding site. Interestingly, the Soret spectral shifts

revealed much lower pKs of 4.4 and 5.6 for wt and K67L Ngb for His64. These differences

indicate a markedly different environment of the His64 imidazole in the equilibrium CO-

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

19

bound form and the five-coordinate deoxy photoproduct species, in which the distal heme

pocket appears quite solvent accessible. In MbCO, the low-pH, protonated species binds CO

faster because removal of His64 from the distal heme pocket upon protonation allows much

easier ligand access to the active site (35). NgbCO shows the opposite pH dependence of λCO.

The slow CO binding to the low-pH Ngb photoproduct species most likely arises from the

better solvent accessibility of the large distal pocket. Water may replace His64 and cause even

more hindrance for exogenous ligands to access the heme iron than with His64 in the charge-

neutral form inside the distal pocket.

Why, then, does the presence of water in the distal heme pocket of H64L Ngb not cause a

substantial decrease in the CO association rate? The H64L Ngb deoxy spectra show

approximately the same amount of five- and six-coordinate heme at room temperature. The

on- and off-rates for water are thus about equal, and the on-rate for water will be much faster

than for CO because [CO] = 1 mM, and [H2O] = 55 M. Both on- and off-rates are thus

expected to be fast compared with CO binding, and the CO on-rate of H64L will be lowered

by only a factor of ~2.

An even more complicated effect of protonation is visible in the pH dependence of the

second kinetic step, N2 (Fig. 9). According to Eq. 4, this quantity depends on both λCO and

λHis (or, more generally, the rate of any endogenous ligand); it reflects the competition

between the exogenous and endogenous ligand. The displacement between the curves at 0.05

and 1 bar CO again shows the bimolecular character of CO binding. However, there is also a

pronounced change in N2 with pH. For native Ngb, N2 has small values at very low (pH 4) and

intermediate pH (8 – 9), and maxima around pH 6 and towards very high pH values.

The steep increase of N2 above pH 9 in wt NgbCO is absent in the K67L mutant. This

suggests that Lys67 acts as a second endogenous ligand at high pH when it becomes

deprotonated. The observed increase of N2 by almost one order of magnitude shows that

Lys67 is the predominant ligand at high pH. This interpretation is consistent with the

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

20

spectroscopic data on the H64L mutant. There, we had assigned a protonation with pK 9.9 to

Lys67. Thus, we conclude that, at alkaline pH, the lysine side chain competes with the

exogenous ligand and the His64 imidazole for the sixth coordination in the native protein. The

second step appearing in the ligand association kinetics of H64L NgbCO at pH > 9 (Fig. 7 c)

supports this line of argumentation. Note that this step is rather short (extending the kinetic

decay by a factor of ~3 at pH 10.6). Whereas the association rate of Lys67 is faster than the

one of His64 (in wt NgbCO) at high pH, the Lys67-Fe bond is so unstable that it persists only

up to 1 ms (in H64L NgbCO) before CO replaces the bound Lys67. Moreover, λ1 of H64L

becomes faster at high pH, which reflects better access of CO to the active site, most likely

because the salt bridge between Lys67 and the heme propionate no longer exists in the neutral

form of the Lys67 side chain.

His64 is responsible for the N2 variations below pH 8. Two protons are attached to the

imidazole sidechain at low pH, whereas the neutral state has two tautomeric forms, one with a

proton at Nε2 and another one with the proton on Nδ1. The first tautomer is known to be more

stable in aqueous solution. However, only the second tautomer can bind to the heme iron by

means of the lone pair on Nε2. At low pH, the distal histidine is most of the time in the

charged conformation outside the distal heme pocket so that it cannot bind to the heme iron.

As a consequence, the rate coefficient λHis and the yield N2 are small. At pH 8, the distal

histidine is predominantly neutral and in the distal pocket in its more stable Nε2 tautomeric

state, which cannot bind to the heme iron. For both wt and K67L NgbCO, the maximum yield

of His64-complexed heme occurs at the titration midpoint (Figs. 8 and 9), which is

characterized by the largest number of transitions between the protonated and unprotonated

forms. We believe that, under the nonequilibrium conditions of the ligand binding

experiments, this may lead to the maximum probability of finding the Nδ1 species in the heme

pocket and thus the highest yield of His64-coordinated heme.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

21

Conclusions

Even though the physiological role of Ngb still awaits further clarification, the binding of

small, gaseous ligands will likely be a key event in its function. In this work, we have used

both steady-state and time-resolved visible spectroscopy to examine the role of distal pocket

residues and their protonation states on ligand binding to ferric and ferrous Ngb. By using

mutants, we were able to assess the detailed roles of His64 and Lys67. His64 clearly provides

significant hindrance to access of exogenous ligands to the active site. Lys67 sequesters the

distal pocket from the solvent environment. A variety of different heme complexes exist in the

ferric and ferrous states, and His64 and Lys67 can both compete as endogenous ligands. The

substantial changes in pK observed for the His64 imidazole sidechain under different

conditions indicates a highly flexible distal pocket, with conformations varying greatly with

respect to solvent accessibility. Most remarkable is the enormous reactivity of the heme iron

towards ligands in Ngb, as reflected in the very large association rates for CO and the ability

of the H64L mutant to bind H2O or OH– at the sixth coordination. The major structural

determinant of the exceptional reactivity is likely the out-of-plane positioning of the iron

towards the distal side (14). A large fraction of bimolecular rebinding indicates that there are

highly efficient ligand escape routes that prevent geminate recombination. Even slight pH

variations cause pronounced changes in the association rate of exogenous ligands near

physiological pH and may thus be relevant to the function of Ngb.

Acknowledgments

We thank Drs. T. Burmester and T. Hankeln (University of Mainz, Germany) for kindly

providing the murine Ngb expression system, Uwe Theilen for constructing and producing the

neuroglobin mutants used in this study and Florian Junginger for technical assistance.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

22

References

1. Hardison, R. C. (1996) Proc. Natl. Acad. Sci. U S A 93, 5675-5679

2. Antonini, E., Brunori, M. (1971) Hemoglobin and Myoglobin in Their Reactions with

Ligands, North-Holland, Amsterdam

3. Trent, J. T., and Hargrove, M. S. (2002) J. Biol. Chem. 277, 19538-19545

4. Pesce, A., Bolognesi, M., Bocedi, A., Ascenzi, P., Dewilde, S., Moens, L., Hankeln,

T., and Burmester, T. (2002) EMBO Rep. 3, 1146-1151

5. Burmester, T., Weich, B., Reinhardt, S., and Hankeln, T. (2000) Nature 407, 520-523

6. Awenius, C., Hankeln, T., and Burmester, T. (2001) Biochem. Biophys. Res. Commun.

287, 418-421

7. Reuss, S., Saaler-Reinhardt, S., Weich, B., Wystub, S., Reuss, M. H., Burmester, T.,

and Hankeln, T. (2002) Neuroscience 115, 645-656

8. Mammen, P. P., Shelton, J. M., Goetsch, S. C., Williams, S. C., Richardson, J. A.,

Garry, M. G., and Garry, D. J. (2002) J. Histochem. Cytochem. 50, 1591-1598

9. Bashford, D., Chothia, C., and Lesk, A. M. (1987) J. Mol. Biol. 196, 199-216

10. Dewilde, S., Kiger, L., Burmester, T., Hankeln, T., Baudin-Creuza, V., Aerts, T.,

Marden, M. C., Caubergs, R., and Moens, L. (2001) J. Biol. Chem. 276, 38949-38955

11. Trent, J. T., Hvitved, A. N., and Hargrove, M. S. (2001) Biochemistry 40, 6155-6163

12. Trent, J. T., Watts, R. A., and Hargrove, M. S. (2001) J. Biol. Chem. 276, 30106-

30110

13. Pesce, A., Dewilde, S., Nardini, M., Moens, L., Ascenzi, P., Hankeln, T., Burmester,

T., and Bolognesi, M. (2003) Structure (Camb) 11, 1087-1095

14. Vallone, B., Nienhaus, K., Brunori, M., and Nienhaus, G. U. (2004) Proteins, in press

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

23

15. Couture, M., Das, T. K., Savard, P. Y., Ouellet, Y., Wittenberg, J. B., Wittenberg, B.

A., Rousseau, D. L., and Guertin, M. (2000) Eur. J. Biochem. 267, 4770-4780

16. Wittenberg, J. B., Bolognesi, M., Wittenberg, B. A., and Guertin, M. (2002) J. Biol.

Chem. 277, 871-874

17. Arrendondo-Peter, R., Hargrove, M. S., Sarath, G., Moran, J. F., Lohrman, J., Olson,

J. S., and Klucas, R. V. (1997) Plant Physiol. 115, 1259-1266

18. Arredondo-Peter, R., Hargrove, M. S., Moran, J. F., Sarath, G., and Klucas, R. V.

(1998) Plant Physiol. 118, 1121-1125

19. Duff, S. M., Wittenberg, J. B., and Hill, R. D. (1997) J. Biol. Chem. 272, 16746-16752

20. Hankeln, T., Jaenicke, V., Kiger, L., Dewilde, S., Ungerechts, G., Schmidt, M., Urban,

J., Marden, M. C., Moens, L., and Burmester, T. (2002) J. Biol. Chem. 277, 29012-

29017

21. Venis, S. (2001) Lancet 358, 2055

22. Sun, Y., Jin, K., Mao, X. O., Zhu, Y., and Greenberg, D. A. (2001) Proc. Natl. Acad.

Sci. U S A 98, 15306-15311

23. Sun, Y., Jin, K., Peel, A., Mao, X. O., Xie, L., and Greenberg, D. A. (2003) Proc.

Natl. Acad. Sci. U S A 100, 3497-3500

24. Zhu, Y., Sun, Y., Jin, K., and Greenberg, D. A. (2002) Blood 100, 2494-2498

25. Wakasugi, K., Nakano, T., and Morishima, I. (2003) J. Biol. Chem. 278, 36505-36512

26. Du, W., Syvitski, R., Dewilde, S., Moens, L., and La Mar, G. N. (2003) J. Am. Chem.

Soc. 125, 8080-8081

27. Kriegl, J. M., Bhattacharyya, A. J., Nienhaus, K., Deng, P., Minkow, O., and

Nienhaus, G. U. (2002) Proc. Natl. Acad. Sci. U S A 99, 7992-7997

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

24

28. Uno, T., Ryu, D., Tsutsumi, H., Tomisugi, Y., Ishikawa, Y., Wilkinson, A. J., Sato,

H., and Hayashi, T. (2003) J. Biol. Chem. 279, 5886-5893

29. Eaton, W. A., Hofrichter, J. (1981) in Methods in Enzymology, Hemoglobins

(Antonini, E. R.-B., L., Chiancone, E., ed) Vol. 76, pp. 175-261

30. Johnson, J. B., Lamb, D. C., Frauenfelder, H., Müller, J. D., McMahon, B., Nienhaus,

G. U., and Young, R. D. (1996) Biophys. J. 71, 1563-1573

31. Nienhaus, G. U., Mourant, J. R., and Frauenfelder, H. (1992) Proc. Natl. Acad. Sci. U

S A 89, 2902-2906.

32. Steinbach, P. J., Ansari, A., Berendzen, J., Braunstein, D., Chu, K., Cowen, B. R.,

Ehrenstein, D., Frauenfelder, H., Johnson, J. B., Lamb, D. C., Luck, S., Mourant, J. R.,

Nienhaus, G. U., Ormos, P., Philipp, R., Xie, A., and Young, R. D. (1991)

Biochemistry 30, 3988-4001

33. Li, T., Quillin, M. L., Phillips, G. N., Jr., and Olson, J. S. (1994) Biochemistry 33,

1433-1446

34. Müller, J. D., McMahon, B. H., Chien, E. Y., Sligar, S. G., and Nienhaus, G. U.

(1999) Biophys. J. 77, 1036-1051

35. Nienhaus, G. U., Müller, J. D., McMahon, B. H., and Frauenfelder, H. (1997) Physica

D 107, 297-311

36. Yang, F., and Phillips, G. N., Jr. (1996) J. Mol. Biol. 256, 762-774.

37. Sage, J. T., Morikis, D., and Champion, P. M. (1991) Biochemistry 30, 1227-1237

38. Sage, J. T., Li, P., and Champion, P. M. (1991) Biochemistry 30, 1237-1247

39. Bhattacharya, S., Sukits, S. F., MacLaughlin, K. L., and Lecomte, J. T. (1997)

Biophys. J. 73, 3230-3240

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

25

40. Boschcov, P. W., Seodel, W. S., Muradian, J., Tominaga, M., Paiva, A. C. M., and

Juliano, L. (1983) Bioorg. Chem. 12, 34-44

41. Quillin, M. L., Arduini, R. M., Olson, J. S., and Phillips, G. N., Jr. (1993) J. Mol. Biol.

234, 140-155

42. Gersonde, K., Sick, H., Overkamp, M., Smith, K. M., and Parish, D. W. (1986) Eur. J.

Biochem. 157, 393-404

43. Peyton, D. H., La Mar, G. N., and Gersonde, K. (1988) Biochim. Biophys. Acta 954,

82-94

44. Lamb, D. C., Prusakov, V., Engler, N., Ostermann, A., Schellenberg, P., Parak, F. G.,

and Nienhaus, G. U. (1998) J. Am .Chem. Soc. 120, 2981-2982

45. Hvitved, A. N., Trent, J. T., Premer, S. A., and Hargrove, M. S. (2001) J. Biol. Chem.

276, 34714-34721

46. Couture, M., Das, T. K., Lee, H. C., Peisach, J., Rousseau, D. L., Wittenberg, B. A.,

Wittenberg, J. B., and Guertin, M. (1999) J. Biol. Chem. 274, 6898-6910

47. Hargrove, M. S., Brucker, E. A., Stec, B., Sarath, G., Arredondo-Peter, R., Klucas, R.

V., Olson, J. S., and Phillips, G. N., Jr. (2000) Structure Fold. Des. 8, 1005-1014

48. Delgado-Nixon, V. M., Gonzalez, G., and Gilles-Gonzalez, M. A. (2000)

Biochemistry 39, 2685-2691

49. Nienhaus, K., Deng, P., Kriegl, J. M., and Nienhaus, G. U. (2003) Biochemistry 42,

9647-9658

50. Lide, D. R. (1994) CRC Handbook of Chemistry and Physics, 75th edition Ed., CRC

Press Inc., Boca Raton, FL

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

26

Tables

Table 1: Peak positions of the absorption spectra of Ngb and mutants H64L, K67L, H64L-

K67L in different ligation and oxidation states for the different pH regions.

Protein Oxidation state

Ligation state pH pK Soret

(nm) Visible bands

(nm)

ferric 6-coordinate (His64 / His96) 8 411 530 / 566

ferrous 6-coordinate (His64 / His96) 8 423 528 / 558

Wild-type Ngba

ferrous 6-coordinate (CO / His96)

8 3.5 4.4 415.6

416.6 533 / 568

ferric 6-coordinate (His64 / His96) 8 412 533 / 560

ferrous 6-coordinate (His64 / His96) 8 424 529 / 559 K67L

ferrous 6-coordinate (CO - His96)

4 8 5.6 416.5

417.2 533 / 568

ferric 6-coordinate 6 8 11

7.3 9.9

403 409 411

500 / 631 530 / 568 532 / 568

ferrous 5/6-coordinate 6 11 9.9 422 / 436

436 530 / 560 / 571

527 / 557 H64L

ferrous 6-coordinate (CO - His96) 8 419 536 /569

ferric 6-coordinate 6 9 11

8.8 (9)

403 407 421

499 / 530 / 632 540 / 571

ferrous 5/6-coordinate 6 11

426 / 436 425 / 437

529 / 550 / 573 528 / 558 / 572

H64L-K67L

ferrous 6-coordinate (CO - His96) 8 419 537 / 564

a Spectra at pH 7 have been published previously (10,12).

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

27

Table 2: Bimolecular rate coefficients λ’CO

(in µM− 1s −1) for rebinding of CO from the

solvent to Ngb and selected heme proteins, determined at 293 K in 0.1 M sodium phosphate

buffer (pH 8) from 0.05 and 1 bar CO dataa.

Protein λ’CO (µM−1s−1)

Ngb 67 ± 6b

Ngb K67L 76 ± 6b

Ngb H64L 237 ± 60c

Ngb H64L/K67L 300 ± 60c

Mb 0.5 ± 0.05d

Mb H64L 26 ± 2e

rice Hb1 7.2f

rice Hb1H73L 170f

a Solubility of CO in water at 293 K and 1 bar CO: ~1 mM (50).

b; c k’ CO = 72 µM−1s−1 and k’ CO = 200 µM−1s−1 in 0.1 M potassium phosphate (pH 7) at 25°

C (10)

d;e k’ CO = 0.51 µM−1s−1 and k’ CO = 26 µM−1s−1 in 0.1 M phosphate / 1 mM EDTA (pH 7) at

20 − 22 °C (33).

f in 0.02 M Tris (pH 7) at 20 ° C (17).

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

28

Figure Legends

Fig. 1. Active site structure of murine Ngb (14). The heme and selected residues are

depicted in dark and light grey for the major and minor heme conformer, respectively.

Fig. 2. pH-dependent spectral changes of Ngb and mutants. (a) Symbols: position of the

Soret peak of wt and K67L NgbCO, plotted as a function of pH. (b) Absorption changes in the

spectra of the H64L mutant. Closed symbols: ferric form at 409 nm, open symbols: ferrous

form at 557 nm. The ordinate is scaled such that it represents the fraction of the deprotonated

species. (c) Absorption changes in the spectra of ferric H64L-K67L. Closed symbols: changes

in the absorption at 403 nm, open symbols: shift of the Soret maximum. Lines represent best

fits to the data using Eq. 1, with Henderson-Hasselbalch relations (n = 1) except for the Soret

shift data in panel c, which obviously is a cooperative transition with n = 3.5.

Fig. 3. pH-dependence of absorption spectra of Ngb mutants H64L and H64L-K67L. (a,

d) Ferric forms, spectra expanded by a factor of 5 from 450 to 700 nm. (b, e) Ferrous forms,

spectra expanded by a factor of 5 from 480 to 700 nm is. (e, f) Second derivatives of the

ferrous spectra in panels b and e at pH 6 (solid line) and 11 (dotted line), spectra expanded by

a factor of 10 from 480 to 700 nm. Arrows indicate increasing pH.

Fig. 4. Trapping of pentacoordinate ferrous Ngb at cryogenic temperatures by

continuous photolysis of ligated Ngb. (a) NgbCO, (b) H64L NgbCO. Solid and dashed lines

represent spectra recorded, respectively, before and after illumination with light from a

frequency-doubled Nd:YAG laser. Open circles: photolysis difference spectra. (c) Spectra of

ferrous H64L measured at 290 K (dotted), 10 K (solid) and at 10 K after illumination

(dashed). Open circles: photolysis difference spectrum. Closed circles: difference spectrum

calculated from the spectra at 290 and 10 K. (d) Second derivatives of the spectra in (c).

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

29

Fig. 5. Contour plot of the difference spectrum of wt Ngb after photolysis. The

measurement was done at ambient temperature with protein dissolved in 0.1 M buffer, pH 4,

equilibrated with 0.05 bar CO. Solid (dotted) lines indicate positive and negative absorbance

differences with respect to the CO-ligated form. Contours are spaced logarithmically. The

vertical line marks the default monitoring wavelength of 436 nm.

Fig. 6. Comparison of bimolecular rebinding of CO in Ngb and Mb. All proteins were

dissolved in phosphate buffer solution, pH 8, and equilibrated with 0.05 bar CO (native Ngb

also equilibrated with 1 bar CO (open circles)). Solid lines represent fits with one or two

exponentials, respectively.

Fig. 7. pH dependence of the flash photolysis kinetics of wt and H64L NgbCO. (a) Ngb

equilibrated with 1 bar. (b) Ngb equilibrated with 0.05 bar CO. (c) Mutant H64L equilibrated

with 0.05 bar CO.

Fig. 8. pH dependence of the apparent rate coefficients for the bimolecular CO

rebinding step of wt Ngb and mutants. The rates from exponential fits to the experimental

data are represented by symbols; solid lines show fits of the Henderson-Hasselbalch relation

to the pH dependence. Dashed lines represent averages of the pH-independent apparent rate

coefficients.

Fig. 9. pH dependence of the amplitudes N2 of the second kinetic step. NgbCO (spheres)

and mutant K67L (squares), equilibrated with 1 bar (solid symbols) and 0.05 bar CO (open

symbols). Lines are drawn to guide the eye.

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

30

Figures

Figure 1

wat

Lys67

Arg66

Fe His96

wat

His64

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

31

Figure 2

4 5 6 7 8 9 10 110.00.20.40.60.81.0

4 5 6 7 8415.6

416.0

416.4

416.8

417.2

7 8 9 10 110.00.20.40.60.81.0

ferricH64L

fract

ion

of a

lk. f

orm

K67L CO

ferrousH64L

ferricH64L-K67L

wt Ngb CO

(c)

(b)

(a)

Wav

elen

gth

(nm

)

pH

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

32

Figure 3

400 450 500 600

0

400 500 600

0.0

0.5

1.0

1.5

2.0400 500 600

400 450 500 600

dA2 /d

λ2

Wavelength (nm)Wavelength (nm)

300

ferrousH64L-K67L

ferricH64L-K67L

ferricH64L

ferrousH64L

700 700

300700 700

Abso

rban

ce

(b)

(c)

(d)

(e)

(f)

(a)

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

33

Figure 4

-0.5

0.0

0.5

1.0

1.5

400 420 440-0.50.00.51.01.52.02.5

-0.5

0.0

0.5

1.0

400 420 440 460-1.0

-0.5

0.0

0.5

diff.

290 Killum.

dark

H64LCO (20 K)

diff.

dark

(b)

ferrousH64L(10 K)

dark

illum.

illum.

diff.

NgbCO(10 K)

(a)

Abso

rban

ce (O

D)

460380

Wavelength (nm)

(c)

(d)

290 K 10 K illum. 10 K dark

d2 A/dλ

2

Wavelength (nm)

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

34

Figure 5

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

35

Figure 6

-7 -6 -5 -4 -3 -2 -1 0 1 2

-2.0

-1.5

-1.0

-0.5

0.0

293 K

Ngb Ngb K67L Ngb H64L Ngb H64L-K67L Mb Mb H64L

log

N(t)

log (t/s)

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

36

Figure 7

-3.0

-2.0

-1.0

0.0

-3.0

-2.0

-1.0

0.0

-7 -6 -5 -4 -3 -2 -1 0 1-2.0

-1.5

-1.0

-0.5

0.0

(b)

(c)

(a)

Ngb1 bar CO

pH 4.0 pH 6.0 pH 7.0 pH 9.0 pH 10.0

log

N(t)

pH 7.9 pH 9.7 pH 10.6

H64L0.05 bar CO

log

N(t)

Ngb, 0.05 bar CO

pH 4.0 pH 6.0 pH 7.0 pH 9.0 pH 10.0

log

N(t)

log (t/s)

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

37

Figure 8

4 5 6 7 8 9 10

20000

40000

60000

80000

250000300000

11

H64L

K67L

H64L-K67L

wt NgbpK = 6.0

pK = 6.3

λ 1 (s-1

)

pH

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

38

Figure 9

4 6 8 100.001

0.01

0.1

wt Ngb, 1 bar CO wt Ngb, 0.05 bar CO K67L, 1 bar CO K67L, 0.05 bar CO

N2

pH

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Karin Nienhaus, Jan M. Kriegl and G. Ulrich Nienhausligand binding

Structural dynamics in the active site of murine neuroglobin and its effects on

published online March 11, 2004J. Biol. Chem. 

  10.1074/jbc.M401561200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on April 7, 2018

http://ww

w.jbc.org/

Dow

nloaded from