We performed spectral diffusion experiments in trehalose-enriched glycerol/buffer-glass on horseradish peroxidase where the heme was replaced by metal-free mesoporphyrin IX, and compared them with the respective behavior in a pure glycerol/buffer-glass (Schlichter et al., J. Chem. Phys. 2000, 112:3045-3050). Trehalose has a significant influence: spectral diffusion broadening speeds up compared to the trehalose-free glass. This speeding up is attributed to a shortening of the correlation time of the frequency fluctuations most probably by preventing water molecules from leaving the protein interior. Superimposed to the frequency fluctuation dynamics is a relaxation dynamics that manifests itself as an aging process in the spectral diffusion broadening. Although the trehalose environment speeds up the fluctuations, it does not have any influence on the relaxation. Both relaxation and fluctuations are governed by power laws in time. The respective exponents do not seem to change with the protein environment. From the spectral dynamics, the mean square displacement in conformation space can be determined. It is governed by anomalous diffusion. The associated frequency correlation time is incredibly long, demonstrating that proteins at low temperatures are truly nonergodic systems.

We performed pressure-tuning hole-burning experiments on a modified cytochrome c protein in a glycerol/buffer glass. The shift and the broadening of the holes were investigated for various frequencies within the inhomogeneous band. On the basis of a simple model, we were able to estimate the interaction range between chromophore and protein. It is approximately 4.5 A. The parameters that enter the model are the compressibility, the static mean-square displacement, the inhomogeneous width, and the average spectral shift per pressure. From this result and from our experiments on pressure-induced denaturing, we conclude that water molecules have to be brought very close to the chromophore during the denaturation process.

We compare the spectral diffusion dynamics of resorufin doped glycerol/H 2 O-and glycerol/ D 2 O-glass with the respective dynamics of a chromoprotein in the same glass at 4.2 K. Spectral diffusion broadening of photochemical holes is measured over almost four orders of magnitude in time. In all samples there are strong aging phenomena. Resorufin in deuterated water/glycerol is reasonably well-described by the two level system ͑TLS͒ model. In the protonated glass, the TLS model does not seem to describe the experiments reasonably well. In the protein sample it totally fails.

The low temperature conformational dynamics of the heme type protein mesoporphyrin-IX-substituted horseradish peroxidase is investigated by spectral diffusion waiting time/ aging experiments. Spectral diffusion broadening is governed by a power law in time. There is a small but significant aging effect. It is assumed that the conformational dynamics of the protein which leads to the spectral broadening of the burnt-in holes is governed by a diffusion type equation. In this case the shape of the spectral diffusion kernel is Gaussian. This model is contrasted with spectral diffusion phenomena as described by the TLS-model ͑TLS, two level system͒.

Previous experimental studies on a modified cytochrome c have shown that optical hole widths have a powerlaw dependence on waiting time. We show that a phenomenological model, which assumes Gaussian random frequency fluctuations whose two-point time-correlation function is a stretched exponential, is consistent with the experimental data.Optical spectroscopy has proven to be a useful technique for probing the dynamics of proteins at low temperatures. 1 In this paper, we are concerned with recent optical hole burning experiments on protoporphyrin IX-substituted cytochrome c in dimethylformamide/glycerol glass, performed by Fritsch et al. 2 at 4 K. In these experiments, the optical transition of the chromophore (protoporphyrin IX) is inhomogeneously broadened and a narrow-band laser is used to selectively excite only a small fraction of the chromophores (those on resonance with the laser). Some of the excited molecules undergo a photochemical reaction, and the photoproduct absorbs light in a spectral region different from the chromophore. Therefore, when the chromophore's absorption line shape is subsequently scanned, a dip or "hole" appears at the frequency of the burning laser.In a waiting time experiment, the width of the hole is measured as a function of the waiting time, t w , between burning and scanning. Typically, the hole width increases with t w as a result of "spectral diffusion". That is, the transition frequency of an individual chromophore is not static in time, but rather fluctuates due to changes in the chromophore's local environment. As the ensemble of chromophores evolves in time, this leads to a broadening of the hole. In particular, this experiment measures the convolution of the initial hole shape with the waiting-time-dependent "spectral diffusion kernel". The latter is the conditional probability density that a chromophore has a transition frequency ν at time t w given that it had transition frequency ν 0 at time 0.In the simplest scenario, both the initial hole and the (spectral diffusion) kernel are Lorentzian functions of frequency, in which case one can simply subtract the initial hole width from the hole width at t w to obtain the width of the kernel. And there is at least one physical model that leads to a Lorentzian kernel: when the chromophore is interacting in a dipolar manner with a collection of point defects 3 whose internal states change with time. A particular realization of this picture, believed to be appropriate for chromophores in glasses, takes the point defects to be two-level systems (TLS) whose energy asymmetries and tunneling matrix elements are widely distributed. [4][5][6][7] In the usual TLS model, it is known that the width of the kernel increases logarithmically with t w , and indeed, this t w dependence has been seen for many different chromophore/glass systems. 6 The protein experiments, when analyzed by assuming a Lorentzian kernel, show a power-law dependence of the width of the kernel on t w , with an exponent of about 1/2, 2 rather than the logarithmic depen...