Introduction

Histidine residues are of profound functional importance to many proteins. Among the key properties of histidines are their ability to change the protonation state at physiological pH and to bind metal ions. For example, the interconversion between different protonation states of the imidazole group is required for catalysis of numerous enzymes that rely on general acid/base catalysis (Fersht 1998). Therefore, studies of the kinetics and thermodynamics of these interconversions provide insight into the dynamics and function of histidine-containing active sites. NMR is the most powerful technique to identify the different protonation states of histidine residues, including the fully protonated form and the two tautomers of the imidazole ring (Mandel 1965; Markley 1975; Pelton et al. 1993; Day et al. 2003; Perez-Canadilllas et al. 2003; Shimahara et al. 2007; Grey et al. 2006). Furthermore, NMR relaxation experiments such as Carr-Purcell-Meiboom-Gil (CPMG) relaxation dispersion experiments can be used to characterize the exchange between these different forms (Hass et al. 2008a). Relaxation dispersion techniques, in particular, are well suited to extract quantitative information about exchange processes on the micro-to-millisecond time scale (Palmer et al. 2001). In principle, this information is also encoded in the shape of the NMR signals (Sudmeier et al. 1980). However, the line shapes of the protein NMR signals usually only provide qualitative information about exchange processes.

Histidine dynamics has been characterized using the resonances of nearby 15N backbone nuclei (Hass et al. 2004; Kovrigin and Loria 2006; Grey et al. 2006). However, monitoring the histidine dynamics through nuclei within the imidazole group itself, such as the imidazole 13C nuclei may offers several advantages compared to 15N backbone measurements, in particular in combination with specific 13C labeling of the imidazole ring. Thus, besides providing (1) direct reporters on histidine side-chain dynamics, (2) the imidazole 13C approach, unlike the backbone 15N based methods, only requires assignment of 1H–13C HSQC spectrum of the 13C labeled histidine residues, as the relatively few signals in the spectrum will reduce complications arising from signal overlap. Furthermore, (3) the shorter 13C pulse length compared to the 15N pulse length allows faster CPMG pulsing and thereby faster exchange processes to be studied by 13C CPMG dispersion.

In principle, all 13C, 15N, and non-exchangeable 1H nuclei of the imidazole group could be used to monitor the interchange of protonation states. The Hε1 and Hδ2 are usually detectable and have been utilized to study histidine protonation using line-shape analysis (Sudmeier et al. 1980). Also 1H relaxation dispersion experiments have been developed but remain less accurate, mainly because strong dipole-dipole couplings to other protons complicate the 1H relaxation dispersion experiments (Ishima and Torchia 2003). For 13C and 15N nuclei such complications are usually negligible. However, the 15N imidazole signals can be line-broadened beyond detection by exchange between protonation states (Pelton et al. 1993). In contrast, the 13C signals of imidazole are less affected by line broadening arising from such processes and are therefore more likely to be detectable than the 15N signals (Sudmeier et al. 1980, 2003). Among the imidazole 13C nuclei, 13Cε1 is the most straightforward nuclei to monitor because the 13Cε1–Hε1 spin system in the imidazole ring is well approximated by an isolated two-spin system analogous to the 15N–H spin systems in the protein backbone. Therefore, pulse schemes normally used for the backbone 15N CPMG experiments (Loria et al. 1999a; Tollinger et al. 2001) can be used also for the CPMG experiments on the 13Cε1–Hε1 spin system, as shown recently by (Kovrigin and Loria 2006).

Here we investigate the use of 13Cε1 CPMG dispersion measurements to monitor micro-millisecond fluctuations in the protonation states of histidine residues in proteins. Rates of histidine protonation-deprotonation reported in the literature are on the order of several to many thousands per second (Eigen 1963; Sudmeier et al. 1980; Hass et al. 2007, 2008a). Therefore, exchange between different protonation states is likely to be in the fast exchange regime. This, however, makes a determination of the involved exchange processes difficult since the resonances of a given nucleus in the different exchanging conformations merge into a single average signal. Therefore, when the interconversions are in the fast-exchange regime on the NMR time scale (k ex > Δω), a simultaneous determination of the chemical shifts and the populations of various states from spin relaxation data alone is not possible, even in the simple two-site case, and independent information must be obtained, e.g., from chemical shift titrations in the case of exchange between protonated and deprotonated histidines (Hass et al. 2004).

Here reduced plastocyanin (PCu) from Anabaena variabilis (A.v.) (Guss et al. 1986; Badsberg et al. 1996; Schmidt et al. 2006) specifically 13C labeled in the histidine imidazole side chains was used as a model system to evaluate the use of 13Cε1 CPMG dispersion measurements. A.v. PCu contains three histidines with different binding and protonation characteristics. Two of the histidines, His92 and His39 are involved in Cu(I) binding (Fig. 1). One of the metal binding imidazoles (His92) is exposed to the surface and protonates at acidic pH, whereas the other (His39) is buried inside the protein and does not protonate. The third histidine (His61) is surface exposed and can be protonated or exist in two deprotonated tautomer forms but is not involved in metal binding. The interchange of protonation states of the two histidines His61 and His92 has been studied in detail using various backbone 15N relaxation methods (Hass et al. 2004, 2007, 2008a) and it was found that the protonation of His92 is a two-site exchange process (Hass et al. 2004, 2007), while the protonation of His61 is a more complex three-site exchange process involving both protonation and tautomerization (Hass et al. 2004; 2008a). Here we investigate to what extend information about these exchange processes can be obtained also from CPMG measurements on the 13Cε1 nuclei of the three histidines in A.v. PCu.

Fig. 1
figure 1

Upper and middle panel: the three histidine residues in A.v. PCu and their protonation states and locations in the crystal structure (Schmidt et al. 2006). The two copper coordinating histidines are His39 and His92, where only His92 protonates. Lower panel: the two different orientations of the His61 imidazole side chain that are stabilized by hydrogen bonding; in one orientation the δ-tautomer is stabilized in the other the ε-tautomer. The isotope labels indicated in the figure correspond to the labeling used in this work

Full size image

Materials and methods

Expression and purification of uniformly 15N-labeled, 13C-His enriched A. variabilis plastocyanin

Uniformly 15N, 13C-His enriched A. variabilis plastocyanin was heterologously produced in Escherichia coli BL21(DE3) from the previously described strain (Hass et al. 2008a) by fed-batch cultivation (Schmidt et al. 2006) with the following modifications. After depletion of the glucose, portions of 0.8 g glucose (from 4 g/l stock solution) were added until the initial NH4Cl was depleted. Subsequently 0.6 g 15NH4Cl (>99% 15N enriched, Spectra Stable Isotope) and 25 mg 13C-histidine, HCl (98% 13C enriched, Cambridge Isotope Laboratories, CLM-2264, U-13C6) were added. After the next oxygen spike 3.9 g 15NH4Cl and 100 mg 13C-histidine, HCl were added and the plastocyanin expression induced by addition of IPTG as previously described (Schmidt et al. 2006; Hass et al. 2008a). After 2 h, cells were harvested and the plastocyanin purified as previously described (Schmidt et al. 2006). Approximately 50 mg per liter of culture of pure plastocyanin with a peak ratio of A 278/A 597 of 1.12 was obtained. Finally, the purified oxidized plastocyanin was reduced with a fivefold molar excess of sodium ascorbate for 1 h at +4°C. Excess ascorbate was removed and the protein exchanged into a 5% D2O/95% H2O 50 mM NaCl solution and concentrated to 1.0 mM by ultrafiltration using a stirred ultrafiltration cell (Amicon) fitted with a 3 kDa cutoff membrane.

NMR samples

All samples contained 1.0 mM 13C-His labeled reduced A.v. PCu in 5% D2O/95% H2O with 50 mM NaCl. Small amounts of sodium ascorbate were added in order to keep the protein reduced. The sample used for the chemical shift titration contained 20 mM sodium acetate buffer and 20 mM sodium phosphate buffer. NMR samples used for CPMG dispersion contained no buffer.

NMR spectroscopy

All NMR experiments were carried out at a field strength of 11.7 T corresponding to resonance frequencies of 500 MHz and 125.7 MHz of 1H and 13C nuclei, respectively, using a Varian Inova spectrometer equipped with a cold probe. The pH dependence of the 13C chemical shifts of the histidine imidazole side-chain was determined from a series of constant time (CT) 13C HSQC spectra recorded at 12 different pH values in the range from pH = 4.0 to pH 8.5. The CT version of the 13C HSQC experiment was chosen to eliminate the effect from the scalar 13C coupling between the 13Cδ2 and 13Cγ (Vuister and Bax 1992). The CT delay was set to 16 ms. The sweep widths were set to 5.0 kHz in both the 13C and 1H dimension. The number of complex data points acquired in the two dimensions was 160 and 512, respectively.

The pH dependences of the chemical shifts were analyzed according to a single titration unaffected by other titrations in the protein:

$$ \delta = \delta_{0} + {\frac{\Updelta \delta }{{1 + 10^{{\left[ {{\text{pH}} - {\text{p}}K_{\text{a}} } \right]}} }}} $$
(1)

where δ0 is the chemical shift of the deprotonated form and δ0 + Δδ is the chemical shift of the protonated form.

The pulse sequences used to monitor the 13C CPMG dispersion were those used for conventional 15N CT-CPMG dispersion experiments (Tollinger et al. 2001). The 13C and 15N channels were interchanged using the “rfchannel” command in VnmrJ software (Varian Inc.). The length of each INEPT (insensitive nuclei enhanced by polarization transfer) and refocusing periods outside the CPMG delay were 2.5 ms assuming a scalar coupling constant \( {\text{J}}_{{^{13} {\text{C}} -^{1} {\text{H}}}} \) of ~200 s−1. The sweep widths were set to 1.76 kHz and 5.0 kHz in the 13C and 1H dimension, respectively. The 13C sweep width was chosen so that overlap between the 13Cε1-signals and the folded 13Cδ2-signals was avoided and the 13C carrier frequency was set at 138.5 ppm. A series of 16 and 512 complex data were recorded in the 13C and 1H dimension, respectively. The total CPMG relaxation time, Δ, was 20 ms. A series of 16 2D-planes were recorded with different CPMG fields, ν CPMG, ranging from 50 to 1,900 Hz. An additional plane with a CPMG relaxation time of 0 ms was also recorded. The CPMG field strength is defined as ν CPMG = 1/(2τ cp), where τ cp is the delay between the 180° pulses in the CPMG period. The B 1 field strength of the 180° 13C CPMG pulses was 9.6 kHz. GARP 13C-decoupling was used during acquisition. Temperature was calibrated against a methanol sample. Effective R 2 rates, R eff2 , were calculated from the signal intensities, I(ν CPMG) and I 0, using R eff2  = Δ−1ln(I 0/I(ν CPMG)) where I 0 is the intensity in the 2D-plane without the CPMG delay. Errors were estimated as the average standard deviation of double measurements made at three different CPMG field strengths.

The CPMG dispersion data were analyzed using a two-site exchange model and assuming fast exchange (k ex > Δω) (Luz and Meiboom 1963)

$$ R_{2}^{\text{eff}} = R_{2}^{0} + \Updelta \omega^{2} p_{\text{A}} p_{\text{B}} {\frac{1}{{k_{\text{ex}} }}}\left[ {1 - {\frac{{4\nu_{\text{CPMG}} }}{{k_{\text{ex}} }}}\tan \,{\text{h}}\left( {{\frac{{k_{\text{ex}} }}{{4\nu_{\text{CPMG}} }}}} \right)} \right] $$
(2)

where R 02 is the effective R 2 rate at infinitely strong CPMG fields, Δω = γB 0Δδ is the difference in frequency between the two conformations, γ is the gyromagnetic ratio, Δδ the difference in chemical shift of a given nucleus in the two exchanging forms, p A and p B are the populations of the two forms, and k ex is the exchange rate.

Results and discussion

Chemical sift titration

Chemical shift information is valuable in CPMG studies of exchange processes. In the case of exchange, titration data provide information about the chemical shift difference, Δδ, of a nucleus in various exchanging states. If these processes are in the slow (k ex < Δω) or intermediate (k ex ≈ Δω) exchange regime the obtained Δδ values provide an independent check of the Δδ values derived from the CPMG dispersion curves together with the exchange rate, k ex, and the populations p A and p B, (see Eq. 2). In the case of fast exchanging processes, such as the exchange between the histidine protonation states studied here (see above) they are indispensible in order to extract k ex from the dispersion curves.

Here, a series of CT-HSQC spectra were recorded at different pH in order to monitor the titration of the 13C chemical shifts. Buffer was added to ensure pH stability and fast protonation equilibria (Hass et al. 2007). At pH 6.5 all six signals from the three histidines are observed (see Fig. 2). At higher pH the 13Cδ2-signal of His61 decreases in intensity and disappears above pH 8 (Fig. 2B) while the 13Cε1-signal moves substantially and broadens slightly at alkaline pH (Fig. 2A). (Note that resonance broadening in the 13C dimension of the CT-HSQC spectrum will only appear as an intensity loss. The 13C linewidth is defined solely by the window functions and remains unchanged throughout the titration series). The pK a value of 7.4 obtained from fitting (Eq. 1) to the titration data agrees with previous values of the pK a of His61 (see Table 1). Clearly the 13Cε1 is in fast exchange between protonated and deprotonated forms of His61 because the titration curve is continous (Fig. 2C). The 13Cδ2–1Hδ2 cross-peak, which is highly sensitive to tautomerization disappears at pH > 8 most likely because of slow (k ex < Δω) or intermediate (k ex ≈ Δω) tautomer exchange that occurs when the imidazole ring of His61 is deprotonated and the buffer system is shifted towards alkaline values (Hass et al. 2008a). The disappearance of this cross-peak at pH > pK a suggests that His61 not only undergoes protonation, but also undergoes conformational exchange when deprotonated, most likely tautomerization, in agreement with previous findings (Hass et al. 2008a). Curiously, the 13Cδ2 resonance of His61 (opposite the 13Cε1 resonance) does not move upon protonation of His61. In contrast, the corresponding signals of His39 and His92 are all affected by the protonation of His92. At acidic pH the 13Cε1–1Hε1 and 13Cδ2–1Hδ2 cross-peaks from His92 disappear because of exchange broadening in the 1H dimension (Fig. 2).

Fig. 2
figure 2

Upper panel: Overlay of 13C CT-HSQC spectral regions showing the imidazole CH groups in A.v. PCu at 3.5°C and 12 different pH values ranging from pH 8.5 (blue) through pH 7.4 (red) and pH 5.5 (yellow) to pH 4.0 (green). A Region showing the 13Cε1 signals; B Region showing the 13Cδ2 signals. The arrows indicate the direction in which the peaks move with increasing H+ concentration; (x) denotes peaks from impurities, most likely from denatured protein formed at extreme pH values. Lower panel: Chemical shift titration curves of C The 13Cε1 carbons and D13Cδ2 carbons, of His39 (●), His61 (⊡), and His92 (&#x2716;). The chemical shift data versus pH were analyzed using Eq. 1. The fitted parameters are given in Table 1

Full size image
Table 1 Parametersa derived from 13C and 1H chemical shift titration data obtained from 13C HSQC spectrab of the three histidine residues in A.v. plastocyanin at 3.5°C
Full size table

CPMG dispersion

The CPMG dispersion techniques can detect exchange processes with exchange rate, k ex, that are within the range of 100–10,000 s−1, depending on the chemical shift difference, Δω between the exchanging site, the population, p A and p B of these sites, and the CPMG field strength ν CPMG = 1/(2τ cp), (Eq. 2). The most critical experimental conditions that influence the exchange rate, and/or the populations p A and p B are the pH, the temperature, and the buffer concentration, all of which can be adjusted to make the exchange process detectable by CPMG dispersion experiments. In particular, protonation rates are likely to be general acid/base catalyzed where the exchange rate increases linearly with the buffer concentration (Hass et al. 2007). If this is the case, as it is in A.v. PCu, the buffer concentration should be kept as low as possible to ensure that protonation occurs sufficiently slowly. Consistent with our previous 15N CPMG studies (Hass et al. 2004, 2008a), the 13C CPMG studies here confirm that below 10°C and at low buffer concentrations (<1 mM) the proton exchange rates in A.v. PCu enter a regime where the exchange between protonation states can be monitored by standard CPMG dispersion experiments (k ex < 5,000 s−1).

The 13C chemical shifts and R eff2 rates of His92 and His39 are both strongly affected by the protonation of His92. Even though His39 does not undergo protonation, its change in 13C chemical shift upon titration of His92 is similar to that of His92 and the dispersion profiles for the two residues are similar (Fig. 3B, D). Indeed, the rates obtained from the dispersion curves of His39 13C nuclei are in excellent agreement with rates obtained previously from the backbone 15N nuclei (Hass et al. 2004, 2008a) as shown in Fig. 4B. The results of the CPMG dispersion experiments can be validated by comparing them with the results of the chemical shift titrations, which provide independent information about the populations and chemical shift differences between protonated and deprotonated forms. The comparison can also establish the nature of the exchange process observed in the CPMG dispersion experiment. Thus, since only the product p A p B|Δδ|2 can be obtained from CPMG dispersion data in the case of fast exchange, we fixed the populations to those obtained from the chemical shift titrations and determine |Δδ| or, vice versa, fixed the chemical shift differences to those obtained from 13C chemical shift titration and determine the populations. In both cases mutually consistent values are obtained for His39 and His92 at pH 6.5 as shown in Table 2. This consistency confirms that the dispersions observed for His92 and His39 are caused by the exchange between protonated and deprotonated His92.

Fig. 3
figure 3

13Cε1 CPMG dispersion curves A His39 (pH 5.9 and 3.5°C), B His39 (pH 6.5 and 8.5°C), C His61 (pH 9.5 and 3.5°C), D His92 (pH 6.5 and 8.5°C)

Full size image
Fig. 4
figure 4

A CPMG dispersion curves of His39 at pH 5.9, without buffer added and at four different temperatures: (⊙) 3.5°C (●) 8.5°C (⊡) 13.5°C (■) 18.5°C; B Arrhenius’ plot showing the agreement between the exchange rates for the protonation of His92 obtained from 13C CPMG dispersion (●) and the exchange rates calculated from previously obtained backbone 15N relaxation data (⊙) (Hass et al. 2004). The calculation of the latter rates requires the pKa value of His92, which depends on the temperature. In contrast to previously published values (Hass et al. 2004) the temperature dependence of pKa is here taken into account (see Table 2, footnote a). The activation energies, EA, derived from the slopes obtained in B for the two data sets (●) and (⊙) are 38 and 49 kJ/mol, respectively, according to the Arrhenius equation, kex∝exp(−EA/RT)

Full size image
Table 2 Parametersa derived from the CPMG dispersion and chemical shift titration data obtained for the three histidine residues in A.v. plastocyanin
Full size table

For His92 itself, the 13C dispersion profiles are hampered by the decreasing signal intensity as the pH value approaches the pK a value. This problem also occurs for His61 (pK a = 7.4) for which the signal disappears between pH 6.5–8 at low buffer concentrations (not shown). At pH 9.5 the His61 signal is clearly visible and shows small but significant dispersion (Fig. 3C). At this pH the dispersion effect here can not be caused by the protonation/deprotonation equilibrium because <1% of His61 is protonated (Hass et al. 2008a). In stead it is caused by the tautomerization of His61, as indicated by the agreement between the exchange rate of (2.8 ± 0.7)102 s−1 obtained here at pH 9.5 and the exchange rate of (2.2 ± 0.2)102 s−1 obtained previously for the tautomerization reaction of His61 under similar conditions using backbone 15N and 1H CPMG dispersion (Hass et al. 2008a). Thus, in the case of A.v. PCu the 13C dispersion experiments information is obtained about (1) protonation and deprotonation of the imidazole ring of His92, (2) tautomerization of the imidazole ring of His61, and (3) a rearrangement of the side chain of His39 or its environment caused by His92 protonation. These three types of processes and monitoring of them using 13C CPMG dispersion are discussed in the following three sections.

Imidazole protonation

Protonation equilibria of histidine residues are common at physiological condition. These equilibria can cause NMR signals to broaden significantly or even lead to complete loss of signals (Sudmeier et al. 1980; Hass et al. 2004, 2008a), dependent on the pH, exchange rate, k ex, and the difference in chemical shift, Δδ. In the case of imidazole exchange processes the 13C CPMG dispersion experiments are useful, because the 13C chemical shifts are affected only moderately by imidazole protonation (the Δδ values being of the order of 1 ppm, corresponding to Δω = 786 rad/s, at 11.7 Tesla) and the 13C signals therefore are unlikely to be affected by severe linebroadening. Thus, in the case of surface exposed histidines where protonation equilibria typically are faster than 1,000 s−1 (Eigen 1963; Sudmeier et al. 1980; Hass et al. 2007, 2008a), the exchange process is in fast exchange (k ex > Δω) with respect to 13C nuclei and the observed signals are therefore relatively narrow. This contrasts the imidazole 15N nuclei, which experience much larger shifts (Pelton et al. 1993) (~80 ppm, corresponding to Δω = 25,000 rad/s at 11.7 Tesla), and the exchange process is therefore in the slow or intermediate regime for typical imidazole protonation exchange rates. This, in turn, results in excessive line broadening and loss of signals at exchange rates where relaxation dispersion experiments are useful. In proton detected 13C experiments proton signal broadening may be a problem, as observed for the 13Cε1 nucleus of His61 and His92 where the line broadening of the protons is larger than the line broadening of 13C. This leads to a severe loss of signal intensity in the 13C CPMG dispersion experiment and makes it difficult to monitor the protonation of the imidazole group at pH ≈ pK a where the population of the protonated and deprotonated forms are of the same magnitude. For His92 a sufficiently good signal-to-noise ratio was obtained only for a relatively slow exchange rate in the absence of buffer, and only at pH values relatively far from the pK a value (pH = 6.5) where the protonated form is only 5% populated (Fig. 3C). The skewed population enhances the signal-to-noise ratio, however, it also reduces the CPMG dispersion (Eq. 2), which therefore becomes difficult to detect. The signal-to-noise ratio can also be improved if, instead, the exchange process is brought into the fast exchange regime by increasing the exchange rate, k ex, and thereby reducing the linebroadening. This can be done by adding buffers or by raising the temperature. However, in addition to a reduced CPMG dispersion, an increased k ex also leads to more linear CPMG dispersion profiles, from which it is difficult to obtain reliable values of k ex and the chemical shift difference Δδ. The same problems were observed for His61. Therefore, reliable dispersion curves reporting on the protonation and deprotonation of His61 were not obtained.

Imidazole tautomerization

For free histidine the ε-tautomer is favored over the δ-tautomer by the ratio 4:1 (Reynolds et al. 1973). In proteins this ratio can be very different, however both tautomer forms are commonly found in proteins. The tautomer equilibrium can give rise to exchange phenomena, as in the case of His61 in A.v. PCu. Thus, His61 is not only affected by protonation but also undergoes a tautomerization where the ε-tautomer is favored by a ratio 9:1 (Hass et al. 2008a). At pH = 9.5 His61 is only affected by the tautomerization while the effect from protonation is insignificant because the protonated form is only ~1% populated. Therefore, at high pH the dispersion curves of His61 (Fig. 3B) can be analyzed according to the two-site model (Eq. 2). Yet the dispersion of the 13Cε1 carbon is relatively small (|Δδ| = 0.4 ppm) suggesting that at least in the case of His61 the 13Cε1 carbon is only moderately sensitive to tautomerization.

In contrast, previous studies show that 13Cδ2 is strongly affected by tautomerization with a chemical shift difference, Δδ, between the two tautomeric forms of up till 10 ppm (Sudmeier et al. 2003). In qualitative agreement with this the 13Cδ2-signal of His61 disappears with increasing exchange rate at high pH (Fig. 2B). The disappearance of the signal at pH ≫ pK a suggests that the process is in slow or intermediate exchange at 11.7 Tesla. Therefore, since the exchange rate is ~2,800 s−1 (Table 2) Δδ must be >3 ppm. The large chemical shift fluctuation cannot be explained by a protonation/deprotonation process because it occurs at pH ≫ pK a where the signal is unaffected by this process (see above). Rather large chemical shift fluctuation suggests a conformational change of the imidazole side chain of His61 in accordance with previous studies (Hass et al. 2008a), which showed that the imidazole ring of His61 tautomerizes (see Fig. 1, lower panel). The 13Cδ2 Δδ value of ~10 ppm reported previously for imidazole tautomerization (Sudmeier et al. 2003) agrees well with the Δδ value larger than 3 ppm suggested by this work. Finally, unlike the 13Cδ2-signal of His61, the 13Cδ2-signals of His39 and His92 show no linebroadening at high pH, which indicates the absence of tautomerization equilibria for these residues at high pH, and shows that the copper binding prevents tautomerization.

Because of the larger titration shift observed for 13Cδ2 compared to 13Cε1, 13Cδ2 is potentially a useful reporter on tautomerization processes, provided that the 13Cδ2-signal can be observed. Unfortunately CPMG dispersion measurements on 13Cδ2 are hampered by the large scalar coupling (~70 Hz) between 13Cδ2 and 13Cγ, although this problem could be solved using selectively labeled histidine with NMR inactive 12C in the γ-position. Methods for achieving this kind of selective labeling have been developed recently (Teilum et al. 2006).

Other exchange processes

Although His39 neither protonates nor tautomerizes, it is clearly affected by chemical exchange (Fig. 3A). Therefore, the pH dependent changes in 13C chemical shift of His39 must be caused by the titration of other groups in its vicinity, in this case His92. Indeed, the exchange rate obtained from the CPMG dispersion curve of His39 is in good agreement with the exchange rate of His92, while Δδ value of His39 also derived from the dispersion curve is in good agreement with the Δδ value obtained from the titration of His39 (see Table 2). Finally, the temperature dependence of the rate agrees with that obtained previously from 15N relaxation data (Fig. 4B).

The changes in chemical shift observed for His39 may arise from a conformational change of the copper site, which results from the protonation of His92 (Guss et al. 1986), or they may be caused by the enhanced positive charge of the metal site of PCu (Hass et al. 2008b). However, irrespective of the origin it is interesting that His39 provides the best CPMG dispersion curves for monitoring the His92 protonation. Most likely, this is because His92 protonation causes less line broadening in the proton dimension of the His39 signal than of the His92 signal, which gives a significantly better signal-to-noise ratio for His39 than for His92.

Conclusions

The imidazole 13Cε1 and 13Cδ2 titration curves provide information about the protonation states of the imidazole ring and confirm the pK a values of His92 and His61 in A.v. PCu found previously (Hass et al. 2004), while the 13Cε1 CPMG dispersion curves provide quantitative information on the rate of protonation and the associated chemical shift changes. The obtained rate constants are in good agreement with those obtained previously from backbone 15N relaxation studies, yet the 13C approach used here is experimentally simpler and more direct than the usual approach that relies on backbone 15N measurements.

In studies of active sites containing functionally active histidines and in studies of metal sites with histidine ligands, information on the conformation and dynamics of histidine side chains is of particular interest. The 13C CPMG dispersion experiments on the imidazole 13Cε1 provide a straightforward method for studying the dynamics within these sites, dynamics that may be directly coupled to the function of the protein. If selective 13C histidine labeling is used, as in the present study, the approach can be applied even to relatively large proteins. TROSY based CPMG dispersion experiments are available (Loria et al. 1999b) and could further improve the applicability to larger proteins, as deuteration of non-exchangeable non-histidine proton sites could improve the sensitivity of the TROSY experiments. Our study here, as well as previous studies (Kovrigin and Loria 2006) show that imidazole 13C CPMG dispersion is a useful and straightforward tool to obtain detailed insight onto the dynamics associated with histidines and can complement or substitute the more established NMR techniques to study histidine dynamics, i.e. CPMG dispersion experiments on backbone 15N and 1H (Hass et al. 2004, 2008a; Kovrigin and Loria 2006; Grey et al. 2006), and imidazole 1H and 15N line shape analyses (Sudmeier et al. 1980; Hass et al. 2008a). Together with these techniques the 13C experiments presented here form a broad suite of experiments that is highly useful for studying functional dynamics of histidine containing active sites in enzymes and other proteins.