Advertisement
JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M505473200 on July 20, 2005

J. Biol. Chem., Vol. 280, Issue 39, 33652-33659, September 30, 2005
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
280/39/33652    most recent
M505473200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by van Thor, J. J.
Right arrow Articles by Sage, J. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Thor, J. J.
Right arrow Articles by Sage, J. T.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Ultrafast and Low Barrier Motions in the Photoreactions of the Green Fluorescent Protein*

Jasper J. van Thor{ddagger}, Georgi Y. Georgiev§1, Michael Towrie¶, and J. Timothy Sage§

From the {ddagger}Laboratory of Molecular Biophysics, University of Oxford, Rex Richards Building, South Parks Road, Oxford OX1 3QU, United Kingdom, §The Department of Physics and Center for Interdisciplinary Research on Complex Systems, Northeastern University, Boston, Massachusetts 02115, and the Central Laser Facility, Council for the Central Laboratory of the Research Councils, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire OX11 0QX, United Kingdom

Received for publication, May 18, 2005 , and in revised form, June 29, 2005.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Green fluorescent protein (GFP) fluoresces efficiently under blue excitation despite major electrostatic rearrangements resulting from photoionization of the chromophore and neutralization of Glu-222. A competing phototransformation process, which ionizes the chromophore and decarboxylates Glu-222, mimics the electrostatic and structural changes in the fluorescence photocycle. Structural and spectroscopic analysis of the cryogenically stabilized photoproduct at 100 K and a structurally annealed intermediate of the phototransformed protein at 170 K reveals distinct structural relaxations involving protein, chromophore, solvent, and photogenerated CO2. Strong structural changes of the 100 K photoproduct after decarboxylation appear exclusively within 15 Å of the chromophore and include the electrostatically driven perturbations of Gln-69, Cys-70, and water molecules in an H-bonding network connecting the chromophore. X-ray crystallography to 1.85 Å resolution and static and picosecond time-resolved IR spectroscopy identify structural mechanisms common to phototransformation and to the fluorescence photocycle. In particular, the appearance of a 1697 cm-1 (+) difference band in both photocycle and phototransformation intermediates is a spectroscopic signature for the structural perturbation of Gln-69. This is taken as evidence for an electrostatically driven dynamic response that is common to both photoreaction pathways. The interactions between the chromophore and the perturbed residues and solvent are decreased or removed in the T203H single and T203H/Q69L double mutants, resulting in a strong reduction of the fluorescence quantum yield. This suggests that the electrostatic response to the transient formation of a buried charge in the wild type is important for the bright fluorescence.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Green fluorescent protein (GFP)2 (1, 2) from Aequorea victoria is highly fluorescent with 400 nm excitation, despite major electrostatic rearrangements resulting from rapid charge transfer to the excited chromophore (3). Rapid excited state proton transfer (ESPT) (4-6) follows excitation of the neutral, phenolic chromophore, and the resulting excited phenolate state I* exhibits high quantum efficiency red-shifted emission at 508 nm, with a 3.0-ns lifetime (3). The mechanism by which the protein environment suppresses non-radiative processes remains unexplained. Proposals for the ESPT pathway include a hydrogen bonding network connecting the chromophore phenolic oxygen to Glu-222 via a water molecule and Ser-205 (7-9). Recently, the possibility of a proton transfer pathway including Glu-222, Asp-82, and Glu-5 was put forward (10). The transient infrared absorption reportedly developing at 1706 cm-1 with excitation of GFP in D2O (9) could not distinguish between these proposals. However, experimental evidence for the identity of the proton acceptor has recently been provided from comparison with the E222D mutant (11).

Low quantum yield electron transfer from Glu-222 to the photoexcited chromophore triggers decarboxylation of the buried carboxyl group (12). This irreversible phototransformation process competes with ESPT that initiates the dominant fluorescence photocycle. The buried carboxylate of Glu-222 stabilizes the protonated chromophore via electrostatic repulsion (13). Neutralization of Glu-222 by light-induced decarboxylation therefore leads to chromophore ionization in equilibrium at pH 8.0 (12). Fig. 1 shows a simplified reaction scheme depicting the major charge redistributions for both the high quantum yield, reversible fluorescence photocycle, and the low quantum yield, irreversible, phototransformation pathway. The protein must relax to accommodate the electrostatic stress generated by chromophore ionization and neutralization of Glu-222, either in the phototransformed protein or following the ESPT reaction in the fluorescence photocycle. In the phototransformation pathway, we refer to the initial "dark" structure with a neutral chromophore ({lambda}max = 400 nm) as GFPA, the initial "lumi" state produced by illumination at 100 K ({lambda}max = 497 nm) as GFPL, the "meta" intermediate generated by thermal annealing of GFPL as GFPM, and the final, fully relaxed photoproduct previously characterized at 300 K (12) as GFPR.

We describe structural events that follow charge transfer to the chromophore at cryogenic temperatures well below the protein dynamical transition (14), or on picosecond timescales at ambient temperature. We find the interior of the rigid {beta}-barrel (13) to be significantly more fluid than the active site of previously characterized photoactive proteins (15-17), supporting the proposed (18) existence of protein conformational dynamics far below the solvent glass transition near 200 K. We conclude that rapid, low barrier structural dynamics stabilize the locally created buried charge and account for the bright fluorescence of GFP.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Crystallization and Phototransformation—Wild type GFP was expressed and purified essentially according to a previous study (13). Crystals were grown at 4 °C in 100 mM Tris/HCl, 50 mM MgCl2, and 15% polyethylene glycol 3350, pH 7.8, and cryo-protected using 20% glycerol, 100 mM Tris/HCl, 50 mM MgCl2, and 15% polyethylene glycol 3350, pH 7.8. For the initial studies, GFPA crystals with various dimensions were exposed to an increasing dosage of short-wavelength UV radiation, and x-ray diffraction data were collected, using an in-house rotating anode source, to 2.4-Å resolution. The progressive disappearance of electron density of the {gamma}-carboxylate of Glu-222 was used as an indicator for phototransformation. For these initial studies structures were not fully refined, and crude phases obtained from molecular replacement, rigid body and limited restrained refinements were sufficient to estimate the electron density changes at Glu-222. The progressive appearance of additional difference features in the unrefined FL - FA map at 2.4-Å resolution resembled those later determined at higher resolution (Fig. 2, A-C), indicating the accumulation of the lumi state GFPL. Off-line microspectrophotometric measurements were done at 100 K to monitor the phototransformation as well as fluorescence and absorption measurements of dissolved crystals after illumination. Together, these tests indicated that with illumination normal to the crystallographic b-axis of a crystal with a typical cross-section of 30 µm x 30 µm, complete phototransformation and trapping of a first intermediate at 100 K was achieved by illumination with a short-wavelength UV source (Universal UV lamp (Type TL-900) CAMAG, Muttenz, Switzerland, {lambda}max = 254 nm), with a dose of 0.1 mJ cm-2 s-1 for 75 min while continuously rotating the crystal. The extinction at 254 nm for half the path length (15 µm) was calculated to be 0.7, using an estimated extinction coefficient of 20 mM-1 cm-1.



View larger version (13K):
[in this window]
[in a new window]
  		FIGURE 1.
Simplified reaction scheme for the fluorescence photocycle and phototransformation reaction pathways emphasizing charge redistribution. Light-induced reactions are indicated with "h{nu}," the electronically excited state is indicated between brackets, including the asterisks, quantum yields are given as "{Phi}," and the thermal and radiative reactions are indicated by unmarked arrows. The quantum yield for the phototransformation reaction is wavelength-dependent (12). Kinetics are not given for the simplified scheme, which does not take into account all known intermediates and associated changes but emphasizes the similarity between the charge redistribution on the chromophore and the buried carboxyl group in the reactant and product states. Photographs of green and yellow solutions of GFP and the decarboxylated photoproduct, respectively, show transmission, whereas the photograph of the fluorescent state shows green emission with UV excitation.

 
X-ray Data Collection—A crystal with dimensions of 30 µm x 30 µm x 200 µm was flash-frozen in liquid nitrogen, mounted in a Cryo-Loop, and maintained at 100 K using an Oxford Instruments Cryojet operating at a flow rate of 6.0 liters/min. X-ray diffraction data were collected at beamline 14.2 at the Synchrotron Radiation Source, Daresbury, UK. The wavelength was 0.978 Å, an Area Detector Systems Corporation Q4R charge-coupled device detector was used, and the front slits were set to 100 µm x 200 µm. Before phototransformation, a first GFPA dataset was recorded to 1.85-Å resolution (TABLE ONE). After phototransformation, the crystal was translated ~50 µm, and a new dataset was recorded for the early photoproduct GFPL (TABLE ONE). Subsequently, the crystal was heated to 170 K with a ramp rate of 6 K/min and kept at this temperature for 15 min for structural annealing. The crystal was re-cooled to 100 K with a ramp rate of 6 K/min, translated by 50 µm, and the final GFPM dataset was recorded (TABLE ONE). The crystal volume exposed to x-ray radiation for more than a single set of images was minimized by the translation, and the beam overlap between datasets was estimated to be <50 µm. The data collection and integration statistics were very similar for all three datasets (TABLE ONE), indicating that no appreciable radiation damage had occurred.


View this table:
[in this window]
[in a new window]
  		TABLE ONE
Data-collection and refinement statistics

 
Structure Solution and Refinement—The intensities were integrated with MOSFLM (19) and scaled with SCALA (20). 5% of the reflections were selected for Rfree calculations, and the same set was selected in the photo-product (FL) and high temperature-annealed (FM) data-sets. Molecular replacement was performed with MOLREP (20), using 1HCJ as a search model. Building was performed using O, and the structures were refined with REFMAC (20) and ARP/WARP (21) without using non-crystallographic symmetry restraints. Refinement of chromophore atoms in the phenolic and imidazolinone rings were done with increased planar restraints. The intensities of the 100 K photoproduct (IL) and the 170 K annealed (IM) diffraction experiments were scaled to the first, dark (IA) dataset using SCALEIT (20), with refinement of scale factors and anisotropic B-factors. IL and IM were scaled to IA with 82,131 and 82,242 common reflections, 1.001 and 1.006 scale factors, R-factors of 0.080 and 0.15, and error-weighted R-factors of 0.078 and 0.11, respectively. Normal probability analysis showed a value of {delta}(real)/{delta}(expected) = 1.3 and 1.4 for centric and acentric reflections for the scaling of IL to IA, {delta}(real)/{delta}(expected) = 1.9 and 2.0 for scaling of IM to IA, and {delta}(real)/{delta}(expected) = 1.2 and 1.4 for scaling of IM to IL. The fitted slopes of the normal probability plots were constant for all resolution bins and had near zero intercepts. The isomorphous differences were distributed normal in all resolutions bins for the scalings performed. FL - FA, FM - FL, and FM - FA difference electron density maps were calculated using FFT (20), with rejection of a (high frequency Fourier term) outlier, and used to aid the model building of the photo-product and annealed structures. The final FL - FA and FM - FL difference maps shown in Figs. 2 and 3C were calculated using phases from the GFPA and the GFPL structures after complete refinement, respectively. For modeling water molecules in the GFPL and thermally annealed GFPM structures that are different from those in the GFPA structure, the criterion applied was that convincing 2Fo - Fc density was present as well as density in Fo - Fc omit maps.

With contouring of the FL - FA difference map at an intensity of 3{sigma}, rather than 5{sigma} (Fig. 2, A-C), additional positive and negative difference electron density is revealed significantly further away from the chromophore, covering a large part of the volume of the protein and its hydration shell. At this contouring level, differences between the four chains become apparent, whereas the very strong signals (>5{sigma}) are more similar in the four different chains

Mutagenesis—The T203H/Q69L double mutant was created starting from an expression construct for PA-GFP (T203H) (22), kindly provided by Jennifer Lippincott-Schwartz, using the QuikChange (Stratagene) protocol for the Q69L mutation using primer sequences CAGCTACGGCGTGCTGTGCTTCAGCCGC and GCGGCTGAAGCACAGCACGCCGTAGCTG.

Static Infrared Measurements—A cryo-loop containing either a hydrated GFP film equilibrated at 33% relative humidity or a GFP solution in 75% glycerol was mounted on an Fourier transform infrared microscope, and data were recorded at 2 cm-1 resolution with the sample temperature controlled by a cold nitrogen gas stream flowing at 5 liters/min. Either the 254 nm line of a mercury lamp or the 413 nm line of a krypton laser illuminated the sample until no further increases in the photogenerated CO2 bands were observed. For room temperature measurements, solutions were mounted between calcium fluoride windows separated by a 15-µm spacer. IR difference spectra recorded with wild type GFP were identical to the GFPuv (F99H/M153T/V163A (23)) solution data presented in Fig. 3. GFPuv was used for most solution studies because of better solubility relative to wild type GFP.

Picosecond Time-resolved Infrared Measurements—Time-resolved IR measurements were recorded at the PIRATE facility at the Rutherford-Appleton Laboratory (24). The 400 nm second harmonic of a Ti:sapphire laser excited a 6-8 mM solution of GFPuv, prepared in either H2O or D2O, and loaded in a 6- to 12-µm path length infrared cell. Difference frequency mixing of the signal and idler of an optical parametric amplifier in a type I AgGaS2 crystal generated a 100-nJ IR pulse with 200-fs pulse width centered at 1720 cm-1 (150 cm-1 full width at half-maximum), which was separated into probe and reference pulses. The IR probe pulse and 400 nm excitation pulse were focused into the sample to ~150-µm diameter spot size. The transmitted probe pulse was dispersed with 4-cm-1 resolution in a grating spectrograph, detected with a 64-element MCT array, and normalized pixel-by-pixel and pulse-by-pulse with respect to the similarly detected reference pulse. Reported spectra are differences between measurements of probe pulses recorded in the presence and absence of the excitation pulse, optically delayed by the given time with respect to the excitation pulse. The instrument response time was 400 fs. Raster scanning and periodic sample replacement avoided photodegradation, as confirmed by monitoring the visible absorption spectra.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Structure of the GFPA Ground State—The space group of the crystals was C2, with four monomers in the asymmetric unit. Structure refinement and building showed differences between the four chains, and hence no NCS restraints were applied at any stage. The asymmetric unit contained two dimers that have an interface that is centered on Phe-223, similar to other dimer structures (12). In the chromophore region of the A structure, in particular with respect to the conformation of Glu-222 and Thr-203, the electron density supports a structure that is similar to the wild type crystal structure 1GFL solved at 1.9-Å resolution at pH 6.8 (13). In particular, the Glu-222 side chain is not in a rotamer conformation, but is best modeled and refined with the angles {chi}1 = 62°, {chi}2 = 153°, and {chi}3 =-145°, although the electron density indicates some disorder of the side chain (11). This conformation contrasts with the wild type structure solved at 2.13-Å resolution at pH 3.9, where Glu-222 was modeled with angles {chi}1 =-29°, {chi}2 =-160°, and {chi}3 =-69° (1EMB (7)). Also, in our GFPA structure, there is no indication of an alternative rotamer conformation of Thr-203, as modeled in the low pH structure (7).



View larger version (38K):
[in this window]
[in a new window]
  		FIGURE 2.
Conformational changes at 100 K (A-C) and under annealing at 170 K (D-F). A, FL - FA electron density difference maps in the asymmetric unit contoured at 5{sigma} (red) and -5{sigma} (blue) reveal features exclusively present in a 15-Å diameter sphere containing the chromophore in each of the four chains. B, negative electron density map for chain A exclusively present at a contour level of -5{sigma}. Strong negative features are found on the OH, O2, O3, O{gamma}2 oxygen atoms as well as the N2, C{alpha}2, C{beta}2, and C{delta}2 atoms of the chromophore. Additional negative electron density features are associated with Thr-203, Ser-205, Phe-64, Val-68, Cys-70, Gln-69, Leu-42, Leu-220, Ser-72, Gln-204, and the crystallographic waters Z78, Z80, Z81, Z111, Z220, Z223, and Z245. C, corresponding positive electron density map, with features associated with Thr-203, Val-224, Leu-42, Gln-204, Ser-72, and the solvent filled cavity. D, FM - FL difference electron density map caused by structural annealing at 170 K contoured at 4{sigma} (orange) and -4{sigma} (cyan). E, detail of the FM-FL difference map in chain A, showing movements of Thr-203 and Gln-204 to perturb Phe-223 in chain D. F, positive difference electron density centered on the cavity-bound crystallographic water molecule Z73 (of GFPA). The isotropic temperature factors for this water, averaged over all four chains, are 13.2 Å2, 13.4 Å2, and 12.0 Å2 in the GFPA, GFPL, and GFPM structures, respectively, corresponding to an average root mean square displacement of 0.41 Å, 0.41 Å, and 0.39 Å.

 
X-ray Structure and IR Spectroscopy of the 100 K GFPL Photoproduct—In contrast with cryogenically trapped intermediates of other photoactive proteins (15-17), unambiguous and strong structural changes at 100 K are evident from the unaveraged FL - FA difference electron density map (Fig. 2, A-C). The normal distribution of isomorphous differences in all resolution bins, the correspondence between R-factors and error-weighted R-factors in the resolution ranges, the fall off of isomorphous differences with increasing resolution, and constant slope of the normal probability plots over all resolution bins all indicate the statistical relevance and low noise level of the strong electron density present in the difference Fourier map. Strong negative features in excess of 20{sigma} cover the O{epsilon}1, O{epsilon}2, and C{delta} atoms of Glu-222 in all four chains, resulting from the specific and complete light-induced decarboxylation of this residue in crystals of GFP. Net charge transfer from Glu-222 to the chromophore drives many additional conformational changes in the chromophore and its immediate environment, and difference density indicates greater active site disorder in GFPL than in GFPA (Fig. 2). The changes include perturbation of the Gln-69 side chain, which moves away from Thr-203 by hinging around the C{alpha} atom. The Cys-70 side chain becomes more disordered, and the distance from the sulfur to the Val-68 carbonyl oxygen increases by 0.15 ± 0.01 Å in the refined coordinates, consistent with the FL - FA difference density (Fig. 2).

Spectroscopic changes also reflect these substantial structural responses to chromophore ionization in both crystals and solutions. The pronounced shift of the visible absorption (Fig. 3B) and prominent spectral changes in the congested 1000-1800 cm-1 region of the GFPL - GFPA IR difference spectrum (Fig. 3A) resemble those previously attributed to ionization of the chromophore (25). In particular, isotope substitutions and density functional theory calculations on a model compound for the GFP chromophore (26) suggest assignment of positive bands associated with vibrations of the phenol group (1497 and 1582 cm-1), the imidazolinone ring (1537 cm-1), and the bridging carbon (1615 cm-1) of the anionic chromophore in the room temperature IR difference spectrum (11, 25).

We identify a new feature having positive and negative peaks at 1697 and 1691 cm-1 with a perturbation of the side chain C=O stretch of Gln-69, based on its characteristic frequency (27) and the absence of this feature in the phototransformed Q69L/T203H mutant (Fig. 4). We found that the Q69L/T203H double mutant can readily be phototransformed, in contrast to the Q69L single mutant, for which the chromophore is ionized in equilibrium (28). Decreased absorbance of a 2552 cm-1 S-H stretching mode (Fig. 3D) is consistent with weakening of the Cys-70 S{gamma}-H... O Val-68 hydrogen bond in GFPL. The appearance of multiple overlapping CO2 bands (Fig. 3C) suggests significant disorder for the photogenerated CO2. Together with the dominant negative contributions to the difference electron density, the additional disorder of the CO2 molecule underscores the unexpected mobility of the chromophore environment at 100 K. Polarized IR data on oriented single crystals3 supports the presence of the Cys-70... Val-68 hydrogen bond and indicates a minimum of three CO2 environments with distinct frequencies and orientations.



View larger version (50K):
[in this window]
[in a new window]
  		FIGURE 3.
Data in A-E characterize differences among the GFPA, GFPL, GFPM, and GFPR structures displayed in F. Difference spectra recorded upon phototransformation of a hydrated film at 100 K, of a solution in 75% glycerol at 200 K, and of an aqueous solution at 293 K reflect GFPL - GFPA, GFPM - GFPA, and GFPR - GFPA differences, respectively. The various spectral regions monitor chromophore and protein side-chain vibrations in the congested 1000-1800 cm-1 region (A), bleaching of the broad absorption band near 400 nm due to the neutral chromophore in the GFPA state and the appearance of red-shifted bands due to the anionic chromophore in the GFPL, GFPM, and GFPR states (B), vibrations of photogenerated CO2 in the 2340 cm-1 region (C), reduced intensity of the Cys-70 S-H stretching vibration at 2552 cm-1 in the GFPL and GFPM states (D), and perturbation of weakly hydrogen-bonded internal water molecules near 2600 cm-1 in a D2O-hydrated GFP film (E). Insets in C show the 2275 cm-1 region for 13CO2 at 100 K and a strong positive feature in the FL - FA difference map near Glu-222 in chain C. Spectra in A, C, D, and E are normalized to maintain constant CO2 band area, except where indicated by explicit scale factors in C. F, coordinates for GFPA (PDB 1W7S [PDB] ), GFPL (PDB 1W7T [PDB] ), GFPM (PDB 1W7U [PDB] ), and GFPR (PDB 1HCJ [PDB] ).

 
Negative signals corresponding to crystallographic waters in the chromophore vicinity also contribute prominently to the FL - FA difference map, accompanied by weaker positive features. These include an incompletely hydrogen-bonded water cluster Z245, Z80, and Z220 (numbering for chain A of GFPA), which initially hydrogen bonds to the side chains of Gln-69, Thr-203, and Glu-222, to the amide nitrogen of Val 68, and to O{gamma}2 of the chromophore (Fig. 2), but undergoes significant restructuring to form a new stable H-bonded position in the fully relaxed GFPR photoproduct state (12). The D2O-hydrated GFPL - GFPA IR difference spectrum (Fig. 3E) reveals signals in a frequency range characteristic for "dangling" O-D bonds. The dipoles of these weakly hydrogen-bonded waters can presumably reorient to stabilize the ionized chromophore more rapidly than in bulk solvent.



View larger version (26K):
[in this window]
[in a new window]
  		FIGURE 4.
Vibrational spectra probe common structural responses to electrostatic rearrangements in the fluorescence and phototransformation pathways. GFPI* - GFPA IR difference spectra, recorded 50 ps after excitation of GFPA with a 400-nm excitation pulse, reveal a 1721-cm-1 band having a frequency and hydrogen isotope sensitivity consistent with the carbonyl stretch of a carboxylic acid that becomes protonated in the GFPI* state. A longer 500-ps delay time is chosen for the D2O solution, to account for the kinetic isotope effect observed for deprotonation of the chromophore (3). The 1721 cm-1 band is absent in static GFPL - GFPA difference spectra recorded at 100 K, where phototransformation eliminates the Glu-222 proton acceptor, but the persistence of the 1697-cm-1 signal suggests a similar electrostatic stress, due to the creation of the anionic chromophore. The absence of the 1697-cm-1 feature in the phototransformation-induced IR difference signal of the Q69L/T203H mutant at 100 K supports assignment of this feature to Gln-69.

 
Solution measurements at 100 K confirmed that selected features of the GFPL - GFPA IR difference spectrum (Fig. 3) develop with identical kinetics, ruling out significant contributions from additional photo-chemical processes at this temperature. Moreover, identical IR spectral changes, including the appearance of CO2 bands, resulted from illumination at either 254 or 413 nm. The latter measurement indicates the absence of significant heating effects, even with the order of magnitude reduction in the phototransformation quantum yield (12) at 413 nm. The total number of photons absorbed during phototransformation of our GFP crystal at 100 K is five orders of magnitude below that reported to enable migration of photolyzed CO in myoglobin at cryogenic temperature, where structural changes of the polypeptide are not observed ina4{sigma} difference map (29). Apparently, significant additional relaxation of GFP does not take place during the few picoseconds needed for the excited chromophore to cool. Bulk heating by the UV light source is discounted on the basis of calculations and experimental tests using a thermocouple device, as is heating from x-ray absorption (30).

X-ray Structure and IR Spectroscopy of the Thermally Annealed GFPM State—The structure evolves further in a GFPL crystal that is heated from 100 to 170 K, maintained at this temperature for 15 min, and re-cooled to 100 K. The resulting FM - FL difference electron density map (Fig. 2D) reveals structural changes throughout the volume covering all the monomers, including the formation of a hydrogen-bonded cluster of new structural waters in the chromophore environment. A prominent series of positive and negative features reflects reorientation of the Thr-203 side chain, coupled to movements of the Gln-204 side chain and the Phe 223 side chain of the neighboring chain in the dimer (Fig. 2E). The new orientation, with Thr-203 within hydrogen bonding distance of the chromophore phenol oxygen, was previously observed in GFPR (12) as well as in the S65T mutant containing an anionic chromophore (31) and contrasts with the small degree of reorientation (estimated at 10% occupancy) in the GFPL structure.

Spectroscopic properties of the GFPM intermediate produced by phototransformation at 200 K ({lambda}max = 480 nm) are intermediate between those observed for the GFPL and GFPR states (Fig. 3, A and B). Relaxation of the chromophore environment following ionization is evident from the evolution of several features with heating to higher temperatures, in particular a phenolate mode at 1362 and 1341 cm-1 at 100 and 293 K, respectively (Fig. 3A). The 1300-1500 cm-1 region of the light-induced GFPM - GFPA difference spectrum recorded at 200 K resembles the GFPR - GFPA difference spectrum recorded at ambient temperature, rather than the GFPL - GFPA spectrum recorded at 100 K (Fig. 3A). However, the persistence of the 1697 cm-1 signal assigned to Gln-69 (Fig. 3A) and the Cys-70 S-H bleach (Fig. 3D) confirms the formation of a distinct intermediate state at 200 K. The IR spectral differences assigned to Gln-69 and to Cys-70 only disappear upon transition to the GFPR state at higher temperatures, where the protein conformation can fully relax to accommodate the molecular stress induced by phototransformation.

The H2O/CO2 cluster exhibits significant mobility in the chromophore cavity upon annealing. The CO2 vibrational signals evolve over the 100-200 K temperature range, coalescing into a single band at 2338 cm-1 at 200 K (Fig. 3C). Although the FM - FA difference map retains the positive feature associated with CO2 in the FL - FA map (Fig. 3C, inset), a new water molecule partially occupies this site in the GFPM structure, whereas no additional or new electron density can be attributed to the displaced CO2 molecule (Fig. 3F). The progressive ordering of solvent in the active site (Fig. 3F) suggests that the GFPL->GFPM and GFPM->GFPR transitions are entropically unfavorable, and thus indicate that the overall pathway is enthalpically driven. Also, strong positive FM - FL density for a well ordered surface-bound crystallographic water Z73 already present in GFPA and GFPL (Fig. 2F) indicates that long range structural annealing enhances the ordering or occupancy of this hydration water.

Picosecond Transient IR Absorption Reveals a Common Structural Response in the Fluorescence and Phototransformation Pathways—The ESPT model for the fluorescence photocycle (7, 8) predicts an intermediate structure I* with an electrostatic environment resembling that following phototransformation, because the chromophore is ionized and the proposed Glu-222 terminal proton acceptor is neutralized. Indeed, time-resolved infrared difference absorption measurements on GFP 50 ps after 400 nm excitation of the neutral chromophore (Fig. 4) reveal a positive feature at 1721 cm-1, in addition to the 1697 cm-1 Gln-69 signal. The 1721 cm-1 frequency, characteristic for C=O stretching of a strongly hydrogen-bonded neutral carboxylic acid (32), and its 11 cm-1 downshift with H/D exchange suggest Glu-222 as the terminal proton acceptor, because other acidic residues are exposed to solvent and are expected to be fully ionized in equilibrium at pH 8. This evidence is in agreement with a recent proposal (9, 11). Further corroboration results from transient infrared measurements on the E222D mutant (11). Both 1721 cm-1 and 1697 cm-1 bands are already observed 5 ps after excitation, signaling very rapid proton transfer to Glu-222 and ultrafast structural perturbation of Gln-69.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
We have investigated the structural and spectroscopic changes that occur in the phototransformation pathway that is triggered by the light-induced decarboxylation of the buried glutamate 222 side chain. Phototransformation at low temperature produces a lumi photoproduct, GFPL, that is characterized by strong structural changes relative to the ground state GFPA, as determined by x-ray crystallography. Spectroscopic changes in the mid-infrared show that the GFPL state is unrelaxed, and difference bands can be assigned to specific amino acid side chains on the basis of the x-ray structure, the polarization of infrared absorption of phototransformed crystals, and mutagenesis. In particular, the 1697 cm-1 is assigned to structural perturbation of Gln-69 and the 2552 cm-1 band to Cys-70. With subsequent heating, structural annealing of the GFPL photoproduct produces the metastable intermediate GFPM, which is characterized by specific solvent reorganization and protein motions. Picosecond time-resolved transient infrared measurements of the fluorescence cycle show that the structural response to optical excitation is similar to the response in the phototransformation reaction pathway.

The presence of the 1697 cm-1 Gln-69 band in the GFPI*, GFPL, and GFPM intermediates points to common structural mechanisms in the fluorescence and phototransformation pathways, and suggests that the structures reported here mimic intermediates in the fluorescence photocycle. This spectroscopic signature reflects the strong structural perturbation of Gln-69 that is evident in the FL - FA difference electron density map (Fig. 2). Structural perturbations of Cys-70, the H-bonded solvent network, and other features that are observed in the FL - FA difference electron map are also very likely to occur during the fluorescence photocycle as part of the electrostatic response to chromophore ionization.

The electronic absorption spectrum of GFPL (Fig. 3B) resembles that of recently identified ground state intermediates I1 (500 nm) and I2 (497 nm) in the fluorescence photocycle at room temperature (33), which relax to GFPA with a 400-ps time constant, although the blue shift of the I1->I2 transition in the fluorescence pathway is smaller in magnitude than the blue shift occurring in the GFPL->GFPM transition in the phototransformation pathway. Complete relaxation to the GFPR structure requires additional structural changes, including relaxation of Gln-69 (Fig. 3A), that are not likely to occur in the fluorescence cycle.

We propose that the very fast, low barrier structural changes reported here are functional in GFP, promoting optimal transformation through the fluorescence photocycle after the initial excited state ionization and inhibiting competing non-radiative processes. These fluid motions of the chromophore environment are unexpected, in light of its proposed role in suppressing the isomerization that quenches the fluorescence of the isolated chromophore (34). Apparently, picosecond timescale structural motions of water molecules and polar side chains in the chromophore environment stabilize the charge redistribution in the radiative state, accounting for the high fluorescence quantum efficiency under blue excitation. The low fluorescence quantum yields of the T203H (22) and T203H/Q69L mutants, where the interactions between the chromophore and the perturbed regions are diminished or removed, further support this proposal. In the T203H/Q69L mutant the chromophore may be partly shielded from solvent in the cavity, but importantly the H-bonding network connecting the chromophore with solvent, Gln-69, and Cys-70 will be disrupted. The spectroscopic properties of the double mutant do not indicate a character change of the electronic transitions, but show non-radiative decay channels to be strongly affected.

Even at low temperature, large changes in local structure of wild type GFP stabilize the charge redistribution following ionization of the buried chromophore. Indeed, excited state deprotonation remains operative at 77 K (at a reduced rate) and fluorescence emission occurs at 504 nm (3). This evidence further supports the existence of low barrier functional dynamics below the solvent glass transition (18). We note that, in the photoactive yellow protein, electrostatic rearrangements resulting from proton transfer between the buried Glu-46 and the chromophore phenolic oxygen ultimately cause the protein to partially unfold (35). This further underscores the dramatic electrostatic rearrangements that are also occurring during the fluorescence cycle of GFP. In GFP, we observe relatively strong difference electron density, in contrast to a number of other light-sensitive protein (15-17). Also, a recent study of the photosynthetic reaction center reported relatively strong difference electron density with illumination at low temperature (36). It should be pointed out that differences in data quality and data scaling cross-correlations hamper absolute comparison with other light-sensitive systems.

Proton transfer and charge redistribution are ubiquitous in biomolecular reactivity (37). The molecular mechanisms that stabilize the electrostatic changes in GFP, which are proposed to be linked to the surprisingly efficient suppression of non-radiative pathways, may therefore have wider ramifications. Structural dynamics are central in the description of quantum catalysis (37, 38) and Marcus electron transfer theory (39, 40). However, these dynamics are often not directly accessible and are generally determined indirectly from kinetic isotope effects or overall rate determinations, for example, in addition to computational approaches. GFP provides a well characterized model system that is suitable to quantify structural dynamics using ultrafast methods.

We expect that improved understanding of the GFP photocycle will facilitate re-engineering the protein to incorporate novel properties (1, 2) without sacrificing the desirable high fluorescence efficiency. In particular, the present results suggest that mutations in the neutral state background, which has the highest known fluorescence quantum efficiency, should retain the hydrogen bonding network, including Gln-69 and Cys-70 (Figs. 2 and 3).


   FOOTNOTES
 
The atomic coordinates and structure factors (codes 1W7S, 1W7SSF, 1W7T, 1W7TSF, 1W7U, and 1W7USF) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

* This work is supported by grants from The Royal Society (University Research Fellowship to J. J. v. T.) and by National Institutes of Health and National Science Foundation Grants GM-52002 and PHY-0240955 (to J. T. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

1 To whom correspondence should be addressed. Tel.: 44-(0)1865-285352; Fax:44-(0)1865-275182; E-mail: jasper{at}biop.ox.ac.uk.

2 The abbreviations used are: GFP, green fluorescent protein; ESPT, excited state proton transfer; GFPL, lumi photoproduct state of GFP; GFPM, metastable intermediate of GFP; GFPR, relaxed photoproduct of GFP. Back

3 G. Y. Georgiev, J. J. van Thor, and J. T. Sage, manuscript in preparation. Back


   ACKNOWLEDGMENTS
 
We thank SRS personnel on beamline ID14-2 for assistance. We acknowledge use of the PIRATE facility at the Rutherford Appleton Laboratory, Chilton, UK (US/22/B/2/04 to J. V. T.) and thank Tony Parker for rapid access. The GFP crystals used in this work were prepared with the assistance of John Hamilton. We thank Keith Moffat for critical reading of the manuscript.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Tsien, R. Y. (1998) Annu. Rev. Biochem. 67, 509-544[CrossRef][Medline] [Order article via Infotrieve]
  2. Lippincott-Schwartz, J., Snapp, E., and Kenworthy, A. (2001) Nat. Rev. Mol. Cell. Biol. 2, 444-456[CrossRef][Medline] [Order article via Infotrieve]
  3. Chattoraj, M., King, B. A., Bublitz, G. U., and Boxer, S. G. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 8362-8367[Abstract/Free Full Text]
  4. Forster, T. (1950) Zeitschrift für Elektrochemie und Angewandte Physikalische Chemie 54, 531-535
  5. Weller, A. (1952) Z. Elektrochem. 56, 662-668
  6. Agmon, N. (2005) J. Phys. Chem. A 109, 13-35[CrossRef][Medline] [Order article via Infotrieve]
  7. Brejc, K., Sixma, T. K., Kitts, P. A., Kain, S. R., Tsien, R. Y., Ormo, M., and Remington, S. J. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 2306-2311[Abstract/Free Full Text]
  8. Palm, G. J., Zdanov, A., Gaitanaris, G. A., Stauber, R., Pavlakis, G. N., and Wlodawer, A. (1997) Nat. Struct. Biol. 4, 361-365[CrossRef][Medline] [Order article via Infotrieve]
  9. Stoner-Ma, D., Jaye, A. A., Matousek, P., Towrie, M., Meech, S. R., and Tonge, P. J. (2005) J. Am. Chem. Soc. 127, 2864-2865[CrossRef][Medline] [Order article via Infotrieve]
  10. Agmon, N. (2005) Biophys. J. 88, 2452-2461[CrossRef][Medline] [Order article via Infotrieve]
  11. van Thor, J. J., Zanetti, G., Ronayne, K., and Towrie M. (2005) J. Phys. Chem. B. 109, 16099-16108[CrossRef][Medline] [Order article via Infotrieve]
  12. van Thor, J. J., Gensch, T., Hellingwerf, K. J., and Johnson, L. N. (2002) Nat. Struct. Biol. 9, 37-41[CrossRef][Medline] [Order article via Infotrieve]
  13. Yang, F., Moss, L. G., and Phillips, G. N., Jr. (1996) Nat. Biotechnol. 14, 1246-1251[CrossRef][Medline] [Order article via Infotrieve]
  14. Vitkup, D., Ringe, D., Petsko, G. A., and Karplus, M. (2000) Nat. Struct. Biol. 7, 34-38[CrossRef][Medline] [Order article via Infotrieve]
  15. Edman, K., Nollert, P., Royant, A., Belrhali, H., Pebay-Peyroula, E., Hajdu, J., Neutze, R., and Landau, E. M. (1999) Nature 401, 822-826[CrossRef][Medline] [Order article via Infotrieve]
  16. Teng, T. Y., Srajer, V., and Moffat, K. (1994) Nat. Struct. Biol. 1, 701-705[CrossRef][Medline] [Order article via Infotrieve]
  17. Schlichting, I., Berendzen, J., Phillips, G. N., Jr., and Sweet, R. M. (1994) Nature 371, 808-812[CrossRef][Medline] [Order article via Infotrieve]
  18. Fenimore, P. W., Frauenfelder, H., McMahon, B. H., and Young, R. D. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 14408-14413[Abstract/Free Full Text]
  19. Leslie, A. G. W. (1992) Joint CCP4 + ESF-EAMCB Newsletter on Protein Crystallography, No. 26, Warrington, United Kingdom
  20. Collaborative Computational Project, Number 4 (1994) Acta Crystallogr. Sect. D Biol. Crystallogr. 50, 760-763[CrossRef][Medline] [Order article via Infotrieve]
  21. Lamzin, V. S., Perrakis, A., Wilson, K.S. (2001) in International Tables for Crystallography. Vol. F: Crystallography of Biological Macromolecules (Rossmann, M. G., and Arnold, E., eds) pp. 720-722, Dordrecht, Kluwer Academic Publishers, The Netherlands
  22. Lippincott-Schwartz, J., and Patterson, G. H. (2003) Science 300, 87-91[Abstract/Free Full Text]
  23. Crameri, A., Whitehorn, E. A., Tate, E., and Stemmer, W. P. (1996) Nat. Biotechnol. 14, 315-319[CrossRef][Medline] [Order article via Infotrieve]
  24. Towrie, M., Grills, D. C., Dyer, J., Weinstein, J. A., Matousek, P., Barton, R., Bailey, P. D., Subramaniam, N., Kwok, W. M., Ma, C., Phillips, D., Parker, A. W., and George, M. W. (2003) Appl. Spectrosc. 57, 367-380[CrossRef][Medline] [Order article via Infotrieve]
  25. van Thor, J. J., Pierik, A. J., Nugteren-Roodzant, I., Xie, A., and Hellingwerf, K. J. (1998) Biochemistry 37, 16915-16921[CrossRef][Medline] [Order article via Infotrieve]
  26. He, X., Bell, A. F., and Tonge, P. J. (2002) Org. Lett. 4, 1523-1526[CrossRef][Medline] [Order article via Infotrieve]
  27. Barth, A. (2000) Prog. Biophys. Mol. Biol. 74, 141-173[CrossRef][Medline] [Order article via Infotrieve]
  28. Delagrave, S., Hawtin, R. E., Silva, C. M., Yang, M. M., and Youvan, D. C. (1995) Biotechnology (N. Y.) 13, 151-154[CrossRef][Medline] [Order article via Infotrieve]
  29. Chu, K., Vojtchovsky, J., McMahon, B. H., Sweet, R. M., Berendzen, J., and Schlichting, I. (2000) Nature 403, 921-923[CrossRef][Medline] [Order article via Infotrieve]
  30. Kriminski, S., Kazmierczak, M., and Thorne, R. E. (2003) Acta Crystallogr. Sect. D Biol. Crystallogr. 59, 697-708[CrossRef][Medline] [Order article via Infotrieve]
  31. Ormo, M., Cubitt, A. B., Kallio, K., Gross, L. A., Tsien, R. Y., and Remington, S. J. (1996) Science 273, 1392-1395[Abstract]
  32. Nie, B., Stutzman, J., and Xie, A. (2005) Biophys. J. 88, 2833-2847[CrossRef][Medline] [Order article via Infotrieve]
  33. Kennis, J. T., Larsen, D. S., van Stokkum, I. H., Vengris, M., van Thor, J. J., and van Grondelle, R. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 17988-17993[Abstract/Free Full Text]
  34. Niwa, H., Inouye, S., Hirano, T., Matsuno, T., Kojima, S., Kubota, M., Ohashi, M., and Tsuji, F. I. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 13617-13622[Abstract/Free Full Text]
  35. Xie, A., Kelemen, L., Hendriks, J., White, B. J., Hellingwerf, K. J., and Hoff, W. D. (2001) Biochemistry 40, 1510-1517[CrossRef][Medline] [Order article via Infotrieve]
  36. Baxter, R. H., Seagle, B. L., Ponomarenko, N., and Norris, J. R. (2004) J. Am. Chem. Soc. 126, 16728-16729[CrossRef][Medline] [Order article via Infotrieve]
  37. Benkovic, S. J., and Hammes-Schiffer, S. (2003) Science 301, 1196-1202[Abstract/Free Full Text]
  38. Sutcliffe, M. J., and Scrutton, N. S. (2002) Eur. J. Biochem. 269, 3096-3102[Medline] [Order article via Infotrieve]
  39. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S., and Dutton, P. L. (1992) Nature 355, 796-802[CrossRef][Medline] [Order article via Infotrieve]
  40. Page, C. C., Moser, C. C., Chen, X., and Dutton, P. L. (1999) Nature 402, 47-52[CrossRef][Medline] [Order article via Infotrieve]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
J.-P. Colletier, D. Bourgeois, B. Sanson, D. Fournier, J. L. Sussman, I. Silman, and M. Weik
Shoot-and-Trap: Use of specific x-ray damage to study structural protein dynamics by temperature-controlled cryo-crystallography
PNAS, August 19, 2008; 105(33): 11742 - 11747.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
280/39/33652    most recent
M505473200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by van Thor, J. J.
Right arrow Articles by Sage, J. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Thor, J. J.
Right arrow Articles by Sage, J. T.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Molecular and Cellular Proteomics 
 Journal of Lipid Research   ASBMB Today 
Copyright © 2005 by the American Society for Biochemistry and Molecular Biology.
Advertisement
spacer
Advertisement
Advertisement