- Email: [email protected]
Steady-State Tryptophan Fluorescence Spectroscopy Study to Probe Tertiary Structure of Proteins in Solid Powders
Steady-State Tryptophan Fluorescence Spectroscopy Study to Probe Tertiary Structure of Proteins in Solid Powders VIKAS K. SHARMA, DEVENDRA S. KALONIA Department of Pharmaceutical Sciences, University of Connecticut, Storrs, Connecticut 06269
Received 20 October 2002; revised 26 November 2002; accepted 2 December 2002
ABSTRACT: The purpose of this work was to obtain information about protein tertiary structure in solid state by using steady state tryptophan (Trp) fluorescence emission spectroscopy on protein powders. Beta-lactoglobulin (bLg) and interferon alpha-2a (IFN) powder samples were studied by fluorescence spectroscopy using a front surface sample holder. Two different sets of dried bLg samples were prepared by vacuum drying of solutions: one containing bLg, and the other containing a mixture of bLg and guanidine hydrochloride. Dried IFN samples were prepared by vacuum drying of IFN solutions and by vacuum drying of polyethylene glycol precipitated IFN. The results obtained from solid samples were compared with the emission scans of these proteins in solutions. The emission scans obtained from protein powders were slightly blue-shifted compared to the solution spectra due to the absence of water. The emission scans were red-shifted for bLg samples dried from solutions containing GuHCl. The magnitude of the shifts in lmax depended on the extent of drying of the samples, which was attributed to the crystallization of GuHCl during the drying process. The shifts in the lmax of the Trp emission spectrum are associated with the changes in the tertiary structure of bLg. In the case of IFN, the emission scans obtained from PEG-precipitated and dried sample were different compared to the emission scans obtained from IFN in solution and from vacuum dried IFN. The double peaks observed in this sample were attributed to the unfolding of the protein. In the presence of trehalose, the two peaks converged to form a single peak, which was similar to solution emission spectra, whereas no change was observed in the presence of mannitol. We conclude that Trp fluorescence spectroscopy provides a simple and reliable means to characterize Trp microenvironment in protein powders that is related to the tertiary conformation of proteins in the solid state. This study shows that the use of fluorescence spectroscopy of proteins can be extended from simple protein aqueous solutions to protein powders, precipitates, and semidried protein samples to gain understanding of protein tertiary structure in these physical states. ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 92:890–899, 2003
Keywords: fluorescence spectroscopy; proteins; tertiary structure; solid state; tryptophan; trehalose; mannitol
INTRODUCTION Recent advances in biotechnology and recombinant DNA technology have led to exponential growth in the commercial production of various proteins and peptides. For therapeutic use, these Correspondence to: Devendra S. Kalonia (Telephone: 860486-3655; Fax: 860-486-4998; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 92, 890–899 (2003) ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association
890
proteins have to be formulated so that the product chemical and physical stability is maintained during its shelf life of at least 1 year to as long as 2 years.1–3 Due to relative instability of proteins in aqueous solutions, the desired shelf life is commonly achieved by formulating proteins as solid powders either by the process of lyophilization4,5 or by spray drying.6,7 The physical stability of proteins during processing as well as during storage in solid state is usually assessed by monitoring changes in the secondary structure or by
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
the extent of aggregation. It is a common practice to investigate the secondary structure of proteins in solid state by second derivative FTIR spectroscopy. Information about the secondary structure elements such as alpha helices, b-sheets, and random coils of protein powders is obtained by analyzing the amide I band of the FTIR spectrum. Several reports exist in the literature, which correlate the denaturation of proteins in solid state to the changes in the second derivative FTIR spectra.8–13 A change in secondary structure is conclusive of the loss in tertiary structure of proteins; however, if no change in secondary structure is observed, one cannot conclude that there is no change in the tertiary structure as well. This is especially true in the case of formation of ‘‘molten globules’’ where the protein tertiary structure may be lost with a minimal change in the secondary structure.14,15 These types of changes may affect long-term storage stability because a loss in tertiary structure opens up the hydrophobic cavity that may further lead to protein aggregation. Currently available techniques to probe protein tertiary structure in the solid state present many restrictions when it comes to the analysis of pharmaceutical protein powders. For example, X-ray diffraction studies can be conducted only on protein crystals, whereas most of the pharmaceutical protein powders have proteins in the amorphous state. Recently, there have been reports on the use of NMR to study protein structures in the solid state.16–19 Although NMR promises to be a good technique to probe solid-state protein structure, one needs to be careful with the issues of radiolabeling proteins as well as for the complete deuterium exchange in the protein molecule. Few attempts have also been made to use phosphorescence spectroscopy20–22 or total internal reflectance fluorescence (TIRF) spectroscopy23 to probe a protein tertiary structure in the dried state. These techniques, however, have been used more on adsorbed protein films and not specifically on pharmaceutical protein powders. Additionally, the use of phosphorescence needs special environmental conditions such as absence of oxygen and an extreme low temperature of 140 K, and hence, requires special equipment. The use of fluorescence spectroscopy has been precluded by these authors on the basis that strong background scattering is observed in the same region where emission scan from proteins is obtained. In this article we report for the first time the use of steady state tryptophan (Trp) fluorescence
891
spectroscopy on pharmaceutical protein powders to monitor changes in the protein tertiary structure without any contribution from background scattering. This technique is simple and similar to the Trp fluorescence spectroscopy of proteins in aqueous solutions. This work evolved as a result of our current efforts to formulate proteins as dry powders using polyethylene glycol (PEG)-induced precipitation followed by vacuum drying. It was observed during our investigations that the protein formed insoluble aggregates during vacuum drying following precipitation by PEG. However, in the presence of trehalose and mannitol, the dried protein was completely soluble. We were interested in investigating the effect of drying on the changes in the tertiary structure of the protein. Steady-state fluorescence spectroscopy of proteins in aqueous solutions has been used widely to obtain information about the changes in the tertiary structure in the microenvironment of Trp.24,25 Intrinsic Trp fluorescence is a function of the polarity of the microenvironment around the Trp and the change in the polarity affects the lmax of Trp fluorescence spectrum. Hence, based on the location of Trp in proteins, the lmax of Trp emission can range from 302 nm (Trp in the hydrophobic interior) to as high as 350 nm (Trp on the surface exposed to water) when excited at 295 nm.24 Unfolding of proteins in aqueous solutions usually results in a change in the microenvironment of Trp from a relatively hydrophobic core to a polar aqueous environment. Therefore, a red shift (shift to a higher wavelength) in the Trp fluorescence emission maximum (lmax ) is usually observed. In other words, a blue shift indicates the presence of hydrophobic environment around the Trp. This understanding of Trp fluorescence emission in solution state, which is well recognized, forms the basis of the present work to probe Trp microenvironment in proteins in the solid state.
MATERIALS AND METHODS Materials All buffer reagents and chemicals used in the present studies were of highest purity grade available from commercial sources and were used without further purification. Polyethylene glycol 1450 was obtained from Dow Chemical Company (Danbury, CT). Guanidine HCl was purchased from Fisher Scientific. Interferon alpha-2a (IFN) was donated generously by Hoffman-La Roche JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
892
SHARMA AND KALONIA
(Nutley, NJ) and was supplied as 1.6-mg/mL solution in pH 5.0 acetate buffer. Beta-lactoglobulin (bLg) was purchased from Sigma (St. Louis, MO). Double distilled water filtered through a 0.22 mm Millipore Millex filter was used for preparation of all solutions. Fluorescence Measurements All fluorescence measurements were carried out using a Perkin-Elmer LS 50 Luminescence Spectrometer. Solution samples were placed in a quartz cuvette for fluorescence measurements. For solid powder samples a special type of front surface sample holder available from PerkinElmer was utilized. This sample holder has a flat fused silica surface on which powder samples, thin films, or highly viscous material can be mounted. For the present studies, solid samples were placed onto a clear Scotch tape (one side adhesive) that was folded at the ends and mounted on the silica surface. The sample holder was positioned such that the incoming excitation beam encounters the solid sample directly and is reflected towards the emission detector. A schematic of this set up is shown in Figure 1(A). Care was taken to cover the tape surface with as much powder as possible. The excitation wavelength was set at 295 nm to minimize contribution from tyrosine fluorescence. The excitation slit width was set at 10 nm and the emission slit width was set at 0.9 nm. The LS-50 spectrometer has fixed slit width settings of 2.5– 15.0 nm adjustable at 0.1 nm increments. It also has a setting labeled as ‘‘0’’. This setting corresponded to a slit width of 0.9 nm and was obtained from the calibration curve of the fluorescence intensity versus known slit width data. A total of three scans were accumulated and averaged to increase signal to noise ratio. All samples were analyzed in triplicate and the variation in the emission maximum was observed to be less than 0.5 nm. Control experiments were carried out to see if the solid samples were affected by exposure to an excitation beam. In repetitive scans no changes were observed in the fluorescence emission with time when continuously exposed to excitation beam for the duration of the experiments. Although solid samples exhibited a high level of scattering at the excitation wavelength, there was no scattering contribution beyond 310 nm either from the tape or from mannitol deposited on tape, as shown in Figure 1(B). For this reason the emission scans from solid samples were collected from 310 to 400 nm. It is clear from JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Figure 1 that there is no significant scattering as well as no fluorescence contribution either from the tape or from mannitol in the wavelength region of interest, and hence, such a system can be conveniently utilized for fluorescence analysis of protein powders. The nonfluorescent properties of mannitol also make it a suitable inert material for analyzing small amounts of protein powders. For solution sample analysis, the buffer scans were subtracted from the sample scans to eliminate the Raman scattering contribution. This was not necessary for solid dried samples. There is no concept of concentration for solid samples;
Figure 1. (A) Schematic of the front surface fused silica cell holder for mounting of solid powder samples. (B) Trp fluorescence emission scans of tape mounted on flat front surface fused silica compared to mannitol deposited on the tape, lexc ¼ 295 nm. Note that there is no scattering contribution above 315 nm.
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
893
therefore, the changes in the fluorescence intensity in the present studies do not serve as a useful parameter to probe protein structural changes. To appropriately compare the wavelength shifts from different samples, all emission scans were normalized to 1.0 with respect to the lmax. All scans were analyzed using Perkin-Elmer’s FLWinlab software. Sample Preparation For solution analysis all protein samples were prepared at concentration of 0.1 mg/mL. Initial studies to obtain Trp emission scan in the solid state were carried out on bLg obtained from Sigma in the lyophilized form. To study the sensitivity of the instrument, bLg was mixed with mannitol in various proportions ranging from 1 to 100% w/w bLg. To prepare a 1% w/w blend, 1.0 mg of bLg was used. The denaturation of bLg was studied using 0, 1.5, 3.0, and 6 M GuHCl. These solution samples were then subjected to vacuum drying at 100 mTorr to obtain semidried and dried samples. The water content of dried samples was measured by Karl Fisher titrimetry. These studies were carried out to investigate the effect of drying on protein solutions in the presence of GuHCl and also to establish the technique for semidried samples. Studies were then carried out on dried IFN samples. IFN dried samples were prepared in two different ways. In the first case IFN was simply vacuum dried out of solution at pH 6.5. In the second case, IFN was first precipitated at pH 6.5 using 25% w/v PEG 1450. The precipitate was collected by centrifugation and subjected to vacuum drying at 100 mTorr at 258C. In a similar set of studies, mannitol and trehalose were added to wet precipitate in ratios ranging from 1:1 w/w to 1:10 w/w of protein:sugar. These samples were then subjected to vacuum drying at 100 mTorr at 258C. Fluorescence studies were performed on IFN solution and dried samples to characterize the Trp microenvironment and, hence, tertiary structure of the protein.
RESULTS AND DISCUSSION Fluorescence Studies on Beta-Lactoglobulin (bLg) Trp fluorescence scans of bLg solution at pH 7.4 and lyophilized bLg are shown in Figure 2(a) and (b). Figure 2(a) is the raw data, whereas Figure 2(b) represents the same data normalized
Figure 2. (A) Trp fluorescence emission scans for bLg solution in acetate buffer, pH 5.0, and lyophilized bLg as obtained from Sigma compared with the background scan from tape and mannitol. (B) Fluorescence emission scans of bLg solution and lyophilized bLg normalized to fluorescence intensity 1.0 at lmax.
to 1.0 at the lmax. Figure 2(a) shows that the fluorescence intensity obtained from lyophilized bLg is much higher compared to that of the tape or just mannitol. The background fluorescence from tape or mannitol does not affect the lmax or the shifts obtained in the lmax. In solution, bLg exhibits a lmax of 336 nm, which indicates that the Trps are relatively in a hydrophobic environment. bLg has two Trps located at positions 19 and 61, and from the three-dimensional structure28 it appears that each of these Trp residues is partially buried with partial exposure to solvent. When lyophilized bLg was subjected to excitation at 295 nm, Trp fluorescence emission was observed with a lmax at 334 nm. It is clear that a good Trp fluorescence emission scan can be obtained from bLg in the solid lyophilized state without any scattering contribution. A blue shift in the lmax in the solid state compared JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
894
SHARMA AND KALONIA
to bLg in solution can be associated with the removal of water in the dried bLg. A 2-nm shift from 336 to 334 nm is significant considering the fact that the emission slit width is set at 0.9 nm and the variation in the lmax for replicate scans of a given sample was not more than 0.5 nm. Because removal of water reduces the polarity around the Trps, the effect is similar to the shift that one will observe if pure Trp would have been transferred from water to a nonpolar solvent such as cyclohexane. However, in the case of bLg, the Trps are already present in a relatively hydrophobic environment; hence, removal of water causes only a small shift in the emission lmax. One of the great features of using this technique is its sensitivity; therefore, small amounts of protein that can be used. This is demonstrated in Figure 3 where fluorescence emission scans were collected for varying amounts of bLg mixed with mannitol. No change in the emission lmax or band shape is observed even at 2% w/w of bLg, which corresponds to 2 mg of the protein in a 100-mg mixture. Only about 10 mg of this mixture is used on the sample holder, which corresponds to 0.2 mg of the protein. This is extremely useful for pharmaceutical proteins that are used in low doses and are often mixed with bulking agents and stabilizers during the process of lyophilization or spray drying. Effect of Guanidine HCl GuHCl is used as a common chemical denaturant for the unfolding of proteins.29–32 Because com-
Figure 3. Normalized Trp emission scans of lyophilized bLg from samples containing different w/ w amounts of the protein diluted with crystalline mannitol. —— 100%, -------50% w/w, – – – – –10% w/ w, ––– 5% w/w, –––– 2% w/w. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
plete unfolding of the protein can be achieved by 6 M GuHCl, it has been used widely to characterize the denatured state of proteins, and hence to compare the unfolded state to the native state. In terms of the Trp fluorescence, unfolding of the protein usually results in a complete exposure of Trps to the solvent, that is, water. This exposure of Trps to water results in a red shift in the lmax of Trp emission scans. Furthermore, because the extent of unfolding depends on the GuHCl concentration, the magnitude of the red shift varies with changes in GuHCl concentration, a higher red shift being observed with higher GuHCl concentration. This is clearly seen in Figure 4, which represents Trp emission scans of bLg in the presence of different concentrations of GuHCl. No change in the emission lmax is observed with 1.5 M GuHCl; however, in the presence of 3 M GuHCl a red shift to 347.5 nm is observed. The maximum red shift occurs with 6 M GuHCl where the lmax is shifted to 350.5 nm. This lmax is usually observed when only Trp is dissolved in buffer. Hence, one can conclude that in the presence of 6.0 M GuHCl, the Trp residues in bLg are completely exposed to water and the protein lies in an unfolded state. We were more interested in investigating the effect of drying on GuHCl containing bLg solutions and their fluorescence scans. For this purpose, GuHCl containing bLg solutions as well as bLg solution at pH 7.4, were subjected to vacuum drying at 100 mTorr. Samples were withdrawn after 10 h of drying and finally after 24 h. At 10 h, none of the samples were completely dried except for the pure bLg solution and the water content varied from 15% w/w to 30% w/w. A higher water
Figure 4. Effect of GuHCl on Trp fluorescence emission scan of bLg in solution.
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
895
content was related to higher GuHCl concentration. Even after 24 h of drying, only samples containing 1.5 M GuHCl were found to have water less than 2% w/w. Samples containing 3 and 6 M GuHCl still contained about 15–20% w/w water. Trp emission scans were collected for all these samples and the results are presented in Figures 5 and 6(a)–(c). Figure 5 shows that the lmax for vacuum-dried bLg appears at 332 nm, which is blue shifted compared to lyophilized bLg. This indicates that the Trps reside in a more hydrophobic environment, suggesting that the tertiary structure is more compact in the vicinity of the Trp residues for vacuum-dried bLg compared to lyophilized bLg. For samples dried from solutions containing GuHCl at different moisture levels, various lmax shifts are observed. The lmax increases initially at 10 h drying for 1.5 M containing GuHCl samples to 343.5 nm and then decreases to 336 nm at final drying [Figure 6(a)], whereas, for the samples containing 3 M [Figure 6(b)] and 6 M [Figure 6(c)] GuHCl the lmax shows blue shift for increasing extent of drying compared to GuHCl containing bLg solutions. This behavior can be explained on the basis of drying and crystallization of GuHCl as drying proceeds. For samples containing 1.5 M GuHCl, initial drying of the samples up to 10 h causes an increase in the concentration of soluble GuHCl, which shifts the protein unfolding equilibrium to the unfolded state. This results in an initial red shift; however, upon complete drying, all of the GuHCl has crystallized out and the protein unfolding equilibrium is now shifted to the native state causing a blue shift. It is to be noted
Figure 6. Effect of vacuum drying on Trp fluorescence emission scans of bLg dried from GuHCl containing bLg solutions (a) 1.5 M GuHCl (b) 3.0 M GuHCl (c) 6.0 M GuHCl.
Figure 5. Trp fluorescence emission scans of bLg solution compared to lyophilized bLg and vacuum dried bLg.
here that bLg dried from 1.5 M containing GuHCl still shows a red shift in lmax compared to lyophilized or vacuum dried bLg indicating some loss in the tertiary structure. The samples containing 3.0 and 6.0 M GuHCl do not dry completely, and the blue shift observed in these cases may be JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
896
SHARMA AND KALONIA
due to a decrease in the amount of soluble GuHCl that results in a low GuHCl:protein ratio, causing a shift in the equilibrium of the protein to the native state. However, the fact that the lmax is still red shifted compared to bLg in solution or dried bLg indicates that the Trps are much more exposed in these samples due to a partial loss of the tertiary structure.
In this section we present results on changes in the tertiary structure of IFN in the solid state during drying. As mentioned earlier in the text, this work actually originated from our efforts to obtain IFN as dried powder by precipitation using PEG and subsequent vacuum drying. When dried in the presence of mannitol or trehalose, IFN was found to be completely soluble in pH 5.0 acetate buffer compared to IFN precipitated in the absence of sugar, which formed insoluble aggregates during drying. We were interested in changes in the tertiary structure of IFN after drying in the absence and presence of trehalose or mannitol. IFN is a small protein that belongs to the family of cytokinines26 and has a MW of 19 kDa. It has two Trps—Trp-70 and Trp-140. IFN is an all-helical protein with about 65% of the structure being present as alpha helices.27 The Trp emission scans of IFN in pH 5.0 buffer and for IFN dried from pH 5.0 buffer are shown in Figure 7. In solution IFN exhibits a lmax of 336 nm, which
indicates that the Trps are in a relatively hydrophobic environment similar to what was observed for bLg. When dried from pH 5.0 buffer the emission scan of IFN is blue shifted with a lmax of 332 nm. As explained before, this blue shift could possibly be due to the absence of water in dried samples, which leaves the Trps in an overall hydrophobic environment compared to IFN in solution. It is important to note that in case of IFN a larger blue shift is observed in the solid state from the solution compared to that observed for bLg. This may be related to the relative position of two Trps in IFN compared to bLg. If the Trps are more exposed to solvent, then removal of water will cause a higher blue shift in solid state, which is what might be happening in case of IFN. Figure 8 represents the Trp emission scans of IFN precipitated by PEG and vacuum dried at 100 mTorr with and without mannitol. The Trp emission scan of wet PEG-precipitated IFN is similar to the emission scan of IFN in solution (data not shown), indicating that in the presence of enough water, PEG itself does not affect the Trp microenvironment. In the absence of mannitol the Trp emission scans of precipitated and vacuum dried IFN, however, exhibited two peaks at 332 and 352 nm, respectively, which is unlike any of the previously observed scans (see Figure 6). The two peaks in the Trp emission scans can result only if the two Trps present in IFN are in entirely different environments and emit independent of each other. This is possible if the protein has partially
Figure 7. Trp fluorescence emission scans of IFN solution at pH 6.5 and vacuum-dried IFN normalized to fluorescence intensity 1.0 at lmax.
Figure 8. Trp fluorescence emission scans of PEGprecipitated and vacuum dried IFN with and without mannitol. The left peak is observed at lmax of 332 nm and the right peak is observed at lmax of 352 nm. 1. IFN:Mannitol 1:10, 2. IFN:Mannitol 1:2, 3. IFN:Mannitol 1:1, 4. No Mannitol.
Fluorescence Studies on Interferon Alpha-2a (IFN)
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
opened up with one Trp still in a relatively hydrophobic environment and the other one lies exposed. Hence, significant change in the tertiary structure of IFN takes place upon drying, most likely in the form of unfolding, which results in this particular shape of the emission scan. An alternative explanation could be that some percentage of the protein molecules has totally lost their tertiary structure, resulting in two peaks—one from the native species, and a red-shifted peak from the denatured species. Upon addition of mannitol to as much as 1:10 w/w protein to mannitol ratio, there is no change in the emission spectra, indicating that mannitol does not protect the tertiary structure of the protein against dehydration stresses. This is usually explained on the basis of crystallization of mannitol during drying. Figure 9 shows Trp emission scans of dried IFN with and without the addition of trehalose. As the trehalose amount is increased in the samples, the second peak observed at lmax of 352 nm reduces to a small bump at 1:1 protein:trehalose ratio and is completely lost at 1:2 w/w and higher protein to trehalose ratio. The loss of the peak at 352 nm indicates that trehalose prevents unfolding of the PEG-precipitated IFN during vacuum drying, thus preserving the microenvironment of the Trps. It is clear from these scans that trehalose protects the protein against unfolding and thus is effective in maintaining the tertiary structure of IFN against the stress of dehydration. This behavior of trehalose is well known, and is attributed to
Figure 9. Trp fluorescence emission scans of PEGprecipitated and vacuum dried IFN with and without Trehalose. The left peak is observed at lmax of 332 nm and the right peak is observed at lmax of 352 nm in scan 4. 1. IFN:Trehalose 1:10, 2. IFN:Trehalose 1:2, 3. IFN:Trehalose 1:1, 4. No Trehalose.
897
its amorphous nature and hydrogen bonding properties.
CONCLUSIONS Protein and peptide drugs are commonly formulated as dry amorphous powders to preserve long-term storage stability. Instability in protein formulations is usually related to the changes in the conformational structure of protein moieties. One of the useful tools to get information about protein conformation in the solid state is FTIR spectroscopy, which provides information about changes in the secondary structure of protein molecules. It is important to note, however, that significant changes in the tertiary structure may take place without much change in the secondary structure elements that may also affect long-term stability of protein therapeutics. These changes in the tertiary structure may open up certain hydrophobic pockets, which are more prone to aggregation. In this article we have reported the use of steady state Trp fluorescence emission spectroscopy to probe Trp microenvironment in protein powders such as bLg and IFN that can be related to overall changes in the tertiary structure of proteins. We observed that with the use of an appropriate sample holder, a good emission scan could be obtained that is representative of the Trp microenvironment. The scans from solid samples is normally blue shifted due to the absence of water in these samples. It is difficult to conclude quantitatively, the extent of changes in the native structure in the solid state compared to the solution; however, one can easily compare the emission scans obtained from various solid samples and qualitatively obtain information about the changes in the tertiary structure. The emission scan with the most blue shift can be assumed to have the most native structure, and the extent of red shift will relate to the extent of changes in the tertiary structure. This is demonstrated both for bLg as well as for IFN. The results obtained with all the samples used in this study are summarized in Table 1. In the case of bLg, different samples prepared and dried from GuHCl containing solutions show different levels of shifts in the lmax that can be directly related to the extent of unfolding of the protein. In the case of IFN, the emission scan obtained from the PEG-precipitated and vacuumdried sample is entirely different when compared to IFN in solution or simply vacuum-dried IFN. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
898
SHARMA AND KALONIA
Table 1. Summary of the Trp Fluorescence Emission lmax Obtained for Various Solution and Dried Samples of bLg and IFN Along with Moisture Content of the Dried Samples Protein bLg
bLg and 1.5 M GuHCl
bLg and 3.0 M GuHCl
bLg and 6.0 M GuHCl
IFN PEG-precipitated IFN
Sample Physical State
Moisture,a % w/w
Solution, pH 7.4 Lyophilized Vacuum dried, no GuHCl Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution Vacuum dried from solution Wet precipitate Vacuum dried Vacuum dried with mannitol Vacuum dried with trehalose
— 1.5 2.1 — 2.5 1.6 — 16.5 10 — 29.5 20 — 1.6 — 2.9 1.5 2.5
lmax, nm 336 334 332 335 343.5 336 347.5 343.5 345.5 350.5 345.5 345.5 336 332 336 332, 352b 332, 352b 332
a
All samples were done in triplicate and the standard deviation was less than 5% of the average. The two values for lmax correspond to the two peaks in the emission scan.
b
This is a clear indication of significant changes in the tertiary structure of protein and also establishes the application of fluorescence spectroscopy to monitor these changes in solid samples. The inability of mannitol to preserve IFN structure and the ability of trehalose to maintain protein structure during drying is further demonstrated in the fluorescence emission scans from these samples. It can be concluded that the use of Trp fluorescence spectroscopy on protein powders can provide important insight into the changes in the tertiary structure pf protein molecules in the solid state. Furthermore, the technique can be extended to proteins in a semisolid or partially dried state such as wet precipitates to probe protein tertiary structure. We hope that this technique provides enormous help to the protein formulation scientists to understand and maintain the long-term stability of protein therapeutics.
ACKNOWLEDGMENTS The authors gratefully acknowledge financial support from the National Science Foundation Industry/University Cooperative Research Center for Pharmaceutical Processing and Hoffman-La JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Roche for the donation of interferon alpha-2a samples.
REFERENCES 1. Manning MC, Patel K, Borchardt RT. 1989. Stability of protein pharmaceuticals. Pharm Res 6:903–918. 2. Arakawa T, Prestrelski S, Kenney WC, Carpenter JF. 1993. Factors affecting short-term and longterm stabilities of proteins. Adv Drug Del Rev 10:1–28. 3. Wang W. 1999. Instability, stabilization, and formulation of liquid protein pharmaceuticals. Int J Pharm 185:129–188. 4. Pikal MJ. 1994. Freeze-drying of proteins: Process, formulation, and stability. In: Cleland JL, Langer R, editors. Formulation and delivery of proteins and peptides. Washington, DC: American Chemical Society, pp 120–133. 5. Wang W. 2000. Lyophilization and development of protein pharmaceuticals. Int J Pharm 203:1–60. 6. Mumenthaler M, Hsu CC, Pearlman R. 1994. Feasibility study on spray-drying protein pharmaceuticals: Recombinant human growth hormone and tissue 0 type plasminogen activator. Pharm Res 11:12–20. 7. Maa YF, Nguyen P-A, Andya JD, Dasovich N, Sweeney TD, Shire SJ, Hsu CC. 1998. Effect of
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
8.
9.
10.
11.
12.
13.
14. 15.
16.
17.
18.
19.
spray drying and subsequent processing conditions on residual moisture content and physical/biochemical stability of protein inhalation powders. Pharm Res 15:768–775. Dong A, Huang P, Caughey WS. 1990. Protein secondary structures in water from second-derivative amide I infrared spectra. Biochemistry 29:3303– 3308. Prestrelski SJ, Tedeschi N, Arakawa T, Carpenter JF. 1993. Dehydration-induced conformational transitions in proteins and their inhibition by stabilizers. Biophys J 65(2):661–671. Prestrelski SJ, Arakawa T, Carpenter JF. 1993. Separation of freezing- and drying-induced denaturation of lyophilized proteins using stress specific stabilization: II. Structural studies using infrared spectroscopy. Arch Biochem Biophys 303:465– 473. Dong A, Caughey WS. 1994. Infrared methods for study of hemoglobin reactions and structures. Methods Enzymol 232:139–175. Griebenow K, Klibanov AM. 1995. Lyophilizationinduced reversible changes in the secondary structure of proteins. Proc Natl Acad Sci USA 92: 10969–10976. Allison SD, Chang B, Randolph TW, Carpenter JF. 1999. Hydrogen bonding between sugar and protein is responsible for inhibition of dehydrationinduced protein unfolding. Arch Biochem Biophys 365:289–298. Ptitsyn OB. 1995. Molten globule and protein folding. Adv Protein Chem 47:83–229. Fink AL. 1995. Compact intermediate states in protein folding. Annu Rev Biophys Biomol Struct 24:495–522. Smith SO, Griffin RG. 1988. High-resolution solidstate NMR of proteins. Annu Rev Phys Chem 39:511–535. Derome AE, Bowden S. 1991. Low-temperature solid-state NMR of proteins. Chem Rev 91:1307– 1320. Fu R, Cross TA. 1999. Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure. Annu Rev Biophys Biomol Struct 28: 235–268. Tycko R. 2001. Biomolecular solid state NMR: Advances in structural methodology and applications to peptide and protein fibrils. Annu Rev Phys Chem 52:575–606.
899
20. Strambini EG, Strambini GB. 2000. Tryptophan phosphorescence as a monitor of protein conformation in molecular films. Biosensors Bioelectron 15:483–490. 21. Gabellieri E, Strambini GB. 2001. Structural perturbations of azurin deposited on solid matrices as revealed by Trp phosphorescence. Biophys J 80:2431–2438. 22. Strambini GB, Gabellieri E. 1984. Intrinsic phosphorescence from proteins in the solid state. Photochem Photobiol 39:725–729. 23. Buijs J, Hlady V. 1997. Adsorption kinetics, conformation, and mobility of the growth hormone and lysozyme on solid surfaces, studied with TIRF. J Colloid Interface Sci 190:171–181. 24. Lakowicz JR. 1982. Principles of fluorescence spectroscopy. New York: Plenum Press. 25. Eftink MR. 1994. The use of fluorescence methods to monitor unfolding transitions in proteins. Biophys J 66:482–501. 26. Haria M, Benfield P. 1995. Interferon a-2a: A review of its pharmacological properties and therapeutic use in the management of viral hepatitis. Drugs 50:873–896. 27. Klaus W, Gsell B, Labhardt AM, Wipf B, Senn H. 1997. The three-dimensional high-resolution structure of human interferon a-2a determined by heteronuclear NMR spectroscopy in solution. J Mol Biol 274:661–675. 28. Uhrinova S, Smith MH, Jameson GB, Uhrin D, Sawyer L, Barlow PN. 2000. Structural changes accompanying pH-induced dissociation of the Blactoglobulin dimer. Biochemistry 39:3565–3574. 29. Tanford C, Kawaharo K, Lapanje S. 1966. Proteins in 6M guanidine hydrochloride. Demonstration of random coil behavior. J Biol Chem 241:1921–1923. 30. Tanford C, Kawaharo K, Lapanje S. 1967. Proteins as random coils. I. Intrinsic viscosities and sedimentation coefficients in concentrated guanidine hydrochloride. J Am Chem Soc 89:729–736. 31. Nozaki Y, Tanford C. 1967. Proteins as random coils. II. Hydrogen ion titration curve of ribonuclease in 6M guanidine hydrochloride. J Am Chem Soc 89:742–749. 32. Tanford C, Kawahara K, Lapanje S, Hooker TM, Zarlengo MH, Salahuddin A, Aune KC, Takagi T. 1967. Proteins as random coils. III. Optical rotatory dispersion in 6M guanidine hydrochloride. J Am Chem Soc 89:5023–5029.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Received 20 October 2002; revised 26 November 2002; accepted 2 December 2002
ABSTRACT: The purpose of this work was to obtain information about protein tertiary structure in solid state by using steady state tryptophan (Trp) fluorescence emission spectroscopy on protein powders. Beta-lactoglobulin (bLg) and interferon alpha-2a (IFN) powder samples were studied by fluorescence spectroscopy using a front surface sample holder. Two different sets of dried bLg samples were prepared by vacuum drying of solutions: one containing bLg, and the other containing a mixture of bLg and guanidine hydrochloride. Dried IFN samples were prepared by vacuum drying of IFN solutions and by vacuum drying of polyethylene glycol precipitated IFN. The results obtained from solid samples were compared with the emission scans of these proteins in solutions. The emission scans obtained from protein powders were slightly blue-shifted compared to the solution spectra due to the absence of water. The emission scans were red-shifted for bLg samples dried from solutions containing GuHCl. The magnitude of the shifts in lmax depended on the extent of drying of the samples, which was attributed to the crystallization of GuHCl during the drying process. The shifts in the lmax of the Trp emission spectrum are associated with the changes in the tertiary structure of bLg. In the case of IFN, the emission scans obtained from PEG-precipitated and dried sample were different compared to the emission scans obtained from IFN in solution and from vacuum dried IFN. The double peaks observed in this sample were attributed to the unfolding of the protein. In the presence of trehalose, the two peaks converged to form a single peak, which was similar to solution emission spectra, whereas no change was observed in the presence of mannitol. We conclude that Trp fluorescence spectroscopy provides a simple and reliable means to characterize Trp microenvironment in protein powders that is related to the tertiary conformation of proteins in the solid state. This study shows that the use of fluorescence spectroscopy of proteins can be extended from simple protein aqueous solutions to protein powders, precipitates, and semidried protein samples to gain understanding of protein tertiary structure in these physical states. ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 92:890–899, 2003
Keywords: fluorescence spectroscopy; proteins; tertiary structure; solid state; tryptophan; trehalose; mannitol
INTRODUCTION Recent advances in biotechnology and recombinant DNA technology have led to exponential growth in the commercial production of various proteins and peptides. For therapeutic use, these Correspondence to: Devendra S. Kalonia (Telephone: 860486-3655; Fax: 860-486-4998; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 92, 890–899 (2003) ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association
890
proteins have to be formulated so that the product chemical and physical stability is maintained during its shelf life of at least 1 year to as long as 2 years.1–3 Due to relative instability of proteins in aqueous solutions, the desired shelf life is commonly achieved by formulating proteins as solid powders either by the process of lyophilization4,5 or by spray drying.6,7 The physical stability of proteins during processing as well as during storage in solid state is usually assessed by monitoring changes in the secondary structure or by
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
the extent of aggregation. It is a common practice to investigate the secondary structure of proteins in solid state by second derivative FTIR spectroscopy. Information about the secondary structure elements such as alpha helices, b-sheets, and random coils of protein powders is obtained by analyzing the amide I band of the FTIR spectrum. Several reports exist in the literature, which correlate the denaturation of proteins in solid state to the changes in the second derivative FTIR spectra.8–13 A change in secondary structure is conclusive of the loss in tertiary structure of proteins; however, if no change in secondary structure is observed, one cannot conclude that there is no change in the tertiary structure as well. This is especially true in the case of formation of ‘‘molten globules’’ where the protein tertiary structure may be lost with a minimal change in the secondary structure.14,15 These types of changes may affect long-term storage stability because a loss in tertiary structure opens up the hydrophobic cavity that may further lead to protein aggregation. Currently available techniques to probe protein tertiary structure in the solid state present many restrictions when it comes to the analysis of pharmaceutical protein powders. For example, X-ray diffraction studies can be conducted only on protein crystals, whereas most of the pharmaceutical protein powders have proteins in the amorphous state. Recently, there have been reports on the use of NMR to study protein structures in the solid state.16–19 Although NMR promises to be a good technique to probe solid-state protein structure, one needs to be careful with the issues of radiolabeling proteins as well as for the complete deuterium exchange in the protein molecule. Few attempts have also been made to use phosphorescence spectroscopy20–22 or total internal reflectance fluorescence (TIRF) spectroscopy23 to probe a protein tertiary structure in the dried state. These techniques, however, have been used more on adsorbed protein films and not specifically on pharmaceutical protein powders. Additionally, the use of phosphorescence needs special environmental conditions such as absence of oxygen and an extreme low temperature of 140 K, and hence, requires special equipment. The use of fluorescence spectroscopy has been precluded by these authors on the basis that strong background scattering is observed in the same region where emission scan from proteins is obtained. In this article we report for the first time the use of steady state tryptophan (Trp) fluorescence
891
spectroscopy on pharmaceutical protein powders to monitor changes in the protein tertiary structure without any contribution from background scattering. This technique is simple and similar to the Trp fluorescence spectroscopy of proteins in aqueous solutions. This work evolved as a result of our current efforts to formulate proteins as dry powders using polyethylene glycol (PEG)-induced precipitation followed by vacuum drying. It was observed during our investigations that the protein formed insoluble aggregates during vacuum drying following precipitation by PEG. However, in the presence of trehalose and mannitol, the dried protein was completely soluble. We were interested in investigating the effect of drying on the changes in the tertiary structure of the protein. Steady-state fluorescence spectroscopy of proteins in aqueous solutions has been used widely to obtain information about the changes in the tertiary structure in the microenvironment of Trp.24,25 Intrinsic Trp fluorescence is a function of the polarity of the microenvironment around the Trp and the change in the polarity affects the lmax of Trp fluorescence spectrum. Hence, based on the location of Trp in proteins, the lmax of Trp emission can range from 302 nm (Trp in the hydrophobic interior) to as high as 350 nm (Trp on the surface exposed to water) when excited at 295 nm.24 Unfolding of proteins in aqueous solutions usually results in a change in the microenvironment of Trp from a relatively hydrophobic core to a polar aqueous environment. Therefore, a red shift (shift to a higher wavelength) in the Trp fluorescence emission maximum (lmax ) is usually observed. In other words, a blue shift indicates the presence of hydrophobic environment around the Trp. This understanding of Trp fluorescence emission in solution state, which is well recognized, forms the basis of the present work to probe Trp microenvironment in proteins in the solid state.
MATERIALS AND METHODS Materials All buffer reagents and chemicals used in the present studies were of highest purity grade available from commercial sources and were used without further purification. Polyethylene glycol 1450 was obtained from Dow Chemical Company (Danbury, CT). Guanidine HCl was purchased from Fisher Scientific. Interferon alpha-2a (IFN) was donated generously by Hoffman-La Roche JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
892
SHARMA AND KALONIA
(Nutley, NJ) and was supplied as 1.6-mg/mL solution in pH 5.0 acetate buffer. Beta-lactoglobulin (bLg) was purchased from Sigma (St. Louis, MO). Double distilled water filtered through a 0.22 mm Millipore Millex filter was used for preparation of all solutions. Fluorescence Measurements All fluorescence measurements were carried out using a Perkin-Elmer LS 50 Luminescence Spectrometer. Solution samples were placed in a quartz cuvette for fluorescence measurements. For solid powder samples a special type of front surface sample holder available from PerkinElmer was utilized. This sample holder has a flat fused silica surface on which powder samples, thin films, or highly viscous material can be mounted. For the present studies, solid samples were placed onto a clear Scotch tape (one side adhesive) that was folded at the ends and mounted on the silica surface. The sample holder was positioned such that the incoming excitation beam encounters the solid sample directly and is reflected towards the emission detector. A schematic of this set up is shown in Figure 1(A). Care was taken to cover the tape surface with as much powder as possible. The excitation wavelength was set at 295 nm to minimize contribution from tyrosine fluorescence. The excitation slit width was set at 10 nm and the emission slit width was set at 0.9 nm. The LS-50 spectrometer has fixed slit width settings of 2.5– 15.0 nm adjustable at 0.1 nm increments. It also has a setting labeled as ‘‘0’’. This setting corresponded to a slit width of 0.9 nm and was obtained from the calibration curve of the fluorescence intensity versus known slit width data. A total of three scans were accumulated and averaged to increase signal to noise ratio. All samples were analyzed in triplicate and the variation in the emission maximum was observed to be less than 0.5 nm. Control experiments were carried out to see if the solid samples were affected by exposure to an excitation beam. In repetitive scans no changes were observed in the fluorescence emission with time when continuously exposed to excitation beam for the duration of the experiments. Although solid samples exhibited a high level of scattering at the excitation wavelength, there was no scattering contribution beyond 310 nm either from the tape or from mannitol deposited on tape, as shown in Figure 1(B). For this reason the emission scans from solid samples were collected from 310 to 400 nm. It is clear from JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Figure 1 that there is no significant scattering as well as no fluorescence contribution either from the tape or from mannitol in the wavelength region of interest, and hence, such a system can be conveniently utilized for fluorescence analysis of protein powders. The nonfluorescent properties of mannitol also make it a suitable inert material for analyzing small amounts of protein powders. For solution sample analysis, the buffer scans were subtracted from the sample scans to eliminate the Raman scattering contribution. This was not necessary for solid dried samples. There is no concept of concentration for solid samples;
Figure 1. (A) Schematic of the front surface fused silica cell holder for mounting of solid powder samples. (B) Trp fluorescence emission scans of tape mounted on flat front surface fused silica compared to mannitol deposited on the tape, lexc ¼ 295 nm. Note that there is no scattering contribution above 315 nm.
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
893
therefore, the changes in the fluorescence intensity in the present studies do not serve as a useful parameter to probe protein structural changes. To appropriately compare the wavelength shifts from different samples, all emission scans were normalized to 1.0 with respect to the lmax. All scans were analyzed using Perkin-Elmer’s FLWinlab software. Sample Preparation For solution analysis all protein samples were prepared at concentration of 0.1 mg/mL. Initial studies to obtain Trp emission scan in the solid state were carried out on bLg obtained from Sigma in the lyophilized form. To study the sensitivity of the instrument, bLg was mixed with mannitol in various proportions ranging from 1 to 100% w/w bLg. To prepare a 1% w/w blend, 1.0 mg of bLg was used. The denaturation of bLg was studied using 0, 1.5, 3.0, and 6 M GuHCl. These solution samples were then subjected to vacuum drying at 100 mTorr to obtain semidried and dried samples. The water content of dried samples was measured by Karl Fisher titrimetry. These studies were carried out to investigate the effect of drying on protein solutions in the presence of GuHCl and also to establish the technique for semidried samples. Studies were then carried out on dried IFN samples. IFN dried samples were prepared in two different ways. In the first case IFN was simply vacuum dried out of solution at pH 6.5. In the second case, IFN was first precipitated at pH 6.5 using 25% w/v PEG 1450. The precipitate was collected by centrifugation and subjected to vacuum drying at 100 mTorr at 258C. In a similar set of studies, mannitol and trehalose were added to wet precipitate in ratios ranging from 1:1 w/w to 1:10 w/w of protein:sugar. These samples were then subjected to vacuum drying at 100 mTorr at 258C. Fluorescence studies were performed on IFN solution and dried samples to characterize the Trp microenvironment and, hence, tertiary structure of the protein.
RESULTS AND DISCUSSION Fluorescence Studies on Beta-Lactoglobulin (bLg) Trp fluorescence scans of bLg solution at pH 7.4 and lyophilized bLg are shown in Figure 2(a) and (b). Figure 2(a) is the raw data, whereas Figure 2(b) represents the same data normalized
Figure 2. (A) Trp fluorescence emission scans for bLg solution in acetate buffer, pH 5.0, and lyophilized bLg as obtained from Sigma compared with the background scan from tape and mannitol. (B) Fluorescence emission scans of bLg solution and lyophilized bLg normalized to fluorescence intensity 1.0 at lmax.
to 1.0 at the lmax. Figure 2(a) shows that the fluorescence intensity obtained from lyophilized bLg is much higher compared to that of the tape or just mannitol. The background fluorescence from tape or mannitol does not affect the lmax or the shifts obtained in the lmax. In solution, bLg exhibits a lmax of 336 nm, which indicates that the Trps are relatively in a hydrophobic environment. bLg has two Trps located at positions 19 and 61, and from the three-dimensional structure28 it appears that each of these Trp residues is partially buried with partial exposure to solvent. When lyophilized bLg was subjected to excitation at 295 nm, Trp fluorescence emission was observed with a lmax at 334 nm. It is clear that a good Trp fluorescence emission scan can be obtained from bLg in the solid lyophilized state without any scattering contribution. A blue shift in the lmax in the solid state compared JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
894
SHARMA AND KALONIA
to bLg in solution can be associated with the removal of water in the dried bLg. A 2-nm shift from 336 to 334 nm is significant considering the fact that the emission slit width is set at 0.9 nm and the variation in the lmax for replicate scans of a given sample was not more than 0.5 nm. Because removal of water reduces the polarity around the Trps, the effect is similar to the shift that one will observe if pure Trp would have been transferred from water to a nonpolar solvent such as cyclohexane. However, in the case of bLg, the Trps are already present in a relatively hydrophobic environment; hence, removal of water causes only a small shift in the emission lmax. One of the great features of using this technique is its sensitivity; therefore, small amounts of protein that can be used. This is demonstrated in Figure 3 where fluorescence emission scans were collected for varying amounts of bLg mixed with mannitol. No change in the emission lmax or band shape is observed even at 2% w/w of bLg, which corresponds to 2 mg of the protein in a 100-mg mixture. Only about 10 mg of this mixture is used on the sample holder, which corresponds to 0.2 mg of the protein. This is extremely useful for pharmaceutical proteins that are used in low doses and are often mixed with bulking agents and stabilizers during the process of lyophilization or spray drying. Effect of Guanidine HCl GuHCl is used as a common chemical denaturant for the unfolding of proteins.29–32 Because com-
Figure 3. Normalized Trp emission scans of lyophilized bLg from samples containing different w/ w amounts of the protein diluted with crystalline mannitol. —— 100%, -------50% w/w, – – – – –10% w/ w, ––– 5% w/w, –––– 2% w/w. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
plete unfolding of the protein can be achieved by 6 M GuHCl, it has been used widely to characterize the denatured state of proteins, and hence to compare the unfolded state to the native state. In terms of the Trp fluorescence, unfolding of the protein usually results in a complete exposure of Trps to the solvent, that is, water. This exposure of Trps to water results in a red shift in the lmax of Trp emission scans. Furthermore, because the extent of unfolding depends on the GuHCl concentration, the magnitude of the red shift varies with changes in GuHCl concentration, a higher red shift being observed with higher GuHCl concentration. This is clearly seen in Figure 4, which represents Trp emission scans of bLg in the presence of different concentrations of GuHCl. No change in the emission lmax is observed with 1.5 M GuHCl; however, in the presence of 3 M GuHCl a red shift to 347.5 nm is observed. The maximum red shift occurs with 6 M GuHCl where the lmax is shifted to 350.5 nm. This lmax is usually observed when only Trp is dissolved in buffer. Hence, one can conclude that in the presence of 6.0 M GuHCl, the Trp residues in bLg are completely exposed to water and the protein lies in an unfolded state. We were more interested in investigating the effect of drying on GuHCl containing bLg solutions and their fluorescence scans. For this purpose, GuHCl containing bLg solutions as well as bLg solution at pH 7.4, were subjected to vacuum drying at 100 mTorr. Samples were withdrawn after 10 h of drying and finally after 24 h. At 10 h, none of the samples were completely dried except for the pure bLg solution and the water content varied from 15% w/w to 30% w/w. A higher water
Figure 4. Effect of GuHCl on Trp fluorescence emission scan of bLg in solution.
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
895
content was related to higher GuHCl concentration. Even after 24 h of drying, only samples containing 1.5 M GuHCl were found to have water less than 2% w/w. Samples containing 3 and 6 M GuHCl still contained about 15–20% w/w water. Trp emission scans were collected for all these samples and the results are presented in Figures 5 and 6(a)–(c). Figure 5 shows that the lmax for vacuum-dried bLg appears at 332 nm, which is blue shifted compared to lyophilized bLg. This indicates that the Trps reside in a more hydrophobic environment, suggesting that the tertiary structure is more compact in the vicinity of the Trp residues for vacuum-dried bLg compared to lyophilized bLg. For samples dried from solutions containing GuHCl at different moisture levels, various lmax shifts are observed. The lmax increases initially at 10 h drying for 1.5 M containing GuHCl samples to 343.5 nm and then decreases to 336 nm at final drying [Figure 6(a)], whereas, for the samples containing 3 M [Figure 6(b)] and 6 M [Figure 6(c)] GuHCl the lmax shows blue shift for increasing extent of drying compared to GuHCl containing bLg solutions. This behavior can be explained on the basis of drying and crystallization of GuHCl as drying proceeds. For samples containing 1.5 M GuHCl, initial drying of the samples up to 10 h causes an increase in the concentration of soluble GuHCl, which shifts the protein unfolding equilibrium to the unfolded state. This results in an initial red shift; however, upon complete drying, all of the GuHCl has crystallized out and the protein unfolding equilibrium is now shifted to the native state causing a blue shift. It is to be noted
Figure 6. Effect of vacuum drying on Trp fluorescence emission scans of bLg dried from GuHCl containing bLg solutions (a) 1.5 M GuHCl (b) 3.0 M GuHCl (c) 6.0 M GuHCl.
Figure 5. Trp fluorescence emission scans of bLg solution compared to lyophilized bLg and vacuum dried bLg.
here that bLg dried from 1.5 M containing GuHCl still shows a red shift in lmax compared to lyophilized or vacuum dried bLg indicating some loss in the tertiary structure. The samples containing 3.0 and 6.0 M GuHCl do not dry completely, and the blue shift observed in these cases may be JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
896
SHARMA AND KALONIA
due to a decrease in the amount of soluble GuHCl that results in a low GuHCl:protein ratio, causing a shift in the equilibrium of the protein to the native state. However, the fact that the lmax is still red shifted compared to bLg in solution or dried bLg indicates that the Trps are much more exposed in these samples due to a partial loss of the tertiary structure.
In this section we present results on changes in the tertiary structure of IFN in the solid state during drying. As mentioned earlier in the text, this work actually originated from our efforts to obtain IFN as dried powder by precipitation using PEG and subsequent vacuum drying. When dried in the presence of mannitol or trehalose, IFN was found to be completely soluble in pH 5.0 acetate buffer compared to IFN precipitated in the absence of sugar, which formed insoluble aggregates during drying. We were interested in changes in the tertiary structure of IFN after drying in the absence and presence of trehalose or mannitol. IFN is a small protein that belongs to the family of cytokinines26 and has a MW of 19 kDa. It has two Trps—Trp-70 and Trp-140. IFN is an all-helical protein with about 65% of the structure being present as alpha helices.27 The Trp emission scans of IFN in pH 5.0 buffer and for IFN dried from pH 5.0 buffer are shown in Figure 7. In solution IFN exhibits a lmax of 336 nm, which
indicates that the Trps are in a relatively hydrophobic environment similar to what was observed for bLg. When dried from pH 5.0 buffer the emission scan of IFN is blue shifted with a lmax of 332 nm. As explained before, this blue shift could possibly be due to the absence of water in dried samples, which leaves the Trps in an overall hydrophobic environment compared to IFN in solution. It is important to note that in case of IFN a larger blue shift is observed in the solid state from the solution compared to that observed for bLg. This may be related to the relative position of two Trps in IFN compared to bLg. If the Trps are more exposed to solvent, then removal of water will cause a higher blue shift in solid state, which is what might be happening in case of IFN. Figure 8 represents the Trp emission scans of IFN precipitated by PEG and vacuum dried at 100 mTorr with and without mannitol. The Trp emission scan of wet PEG-precipitated IFN is similar to the emission scan of IFN in solution (data not shown), indicating that in the presence of enough water, PEG itself does not affect the Trp microenvironment. In the absence of mannitol the Trp emission scans of precipitated and vacuum dried IFN, however, exhibited two peaks at 332 and 352 nm, respectively, which is unlike any of the previously observed scans (see Figure 6). The two peaks in the Trp emission scans can result only if the two Trps present in IFN are in entirely different environments and emit independent of each other. This is possible if the protein has partially
Figure 7. Trp fluorescence emission scans of IFN solution at pH 6.5 and vacuum-dried IFN normalized to fluorescence intensity 1.0 at lmax.
Figure 8. Trp fluorescence emission scans of PEGprecipitated and vacuum dried IFN with and without mannitol. The left peak is observed at lmax of 332 nm and the right peak is observed at lmax of 352 nm. 1. IFN:Mannitol 1:10, 2. IFN:Mannitol 1:2, 3. IFN:Mannitol 1:1, 4. No Mannitol.
Fluorescence Studies on Interferon Alpha-2a (IFN)
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
opened up with one Trp still in a relatively hydrophobic environment and the other one lies exposed. Hence, significant change in the tertiary structure of IFN takes place upon drying, most likely in the form of unfolding, which results in this particular shape of the emission scan. An alternative explanation could be that some percentage of the protein molecules has totally lost their tertiary structure, resulting in two peaks—one from the native species, and a red-shifted peak from the denatured species. Upon addition of mannitol to as much as 1:10 w/w protein to mannitol ratio, there is no change in the emission spectra, indicating that mannitol does not protect the tertiary structure of the protein against dehydration stresses. This is usually explained on the basis of crystallization of mannitol during drying. Figure 9 shows Trp emission scans of dried IFN with and without the addition of trehalose. As the trehalose amount is increased in the samples, the second peak observed at lmax of 352 nm reduces to a small bump at 1:1 protein:trehalose ratio and is completely lost at 1:2 w/w and higher protein to trehalose ratio. The loss of the peak at 352 nm indicates that trehalose prevents unfolding of the PEG-precipitated IFN during vacuum drying, thus preserving the microenvironment of the Trps. It is clear from these scans that trehalose protects the protein against unfolding and thus is effective in maintaining the tertiary structure of IFN against the stress of dehydration. This behavior of trehalose is well known, and is attributed to
Figure 9. Trp fluorescence emission scans of PEGprecipitated and vacuum dried IFN with and without Trehalose. The left peak is observed at lmax of 332 nm and the right peak is observed at lmax of 352 nm in scan 4. 1. IFN:Trehalose 1:10, 2. IFN:Trehalose 1:2, 3. IFN:Trehalose 1:1, 4. No Trehalose.
897
its amorphous nature and hydrogen bonding properties.
CONCLUSIONS Protein and peptide drugs are commonly formulated as dry amorphous powders to preserve long-term storage stability. Instability in protein formulations is usually related to the changes in the conformational structure of protein moieties. One of the useful tools to get information about protein conformation in the solid state is FTIR spectroscopy, which provides information about changes in the secondary structure of protein molecules. It is important to note, however, that significant changes in the tertiary structure may take place without much change in the secondary structure elements that may also affect long-term stability of protein therapeutics. These changes in the tertiary structure may open up certain hydrophobic pockets, which are more prone to aggregation. In this article we have reported the use of steady state Trp fluorescence emission spectroscopy to probe Trp microenvironment in protein powders such as bLg and IFN that can be related to overall changes in the tertiary structure of proteins. We observed that with the use of an appropriate sample holder, a good emission scan could be obtained that is representative of the Trp microenvironment. The scans from solid samples is normally blue shifted due to the absence of water in these samples. It is difficult to conclude quantitatively, the extent of changes in the native structure in the solid state compared to the solution; however, one can easily compare the emission scans obtained from various solid samples and qualitatively obtain information about the changes in the tertiary structure. The emission scan with the most blue shift can be assumed to have the most native structure, and the extent of red shift will relate to the extent of changes in the tertiary structure. This is demonstrated both for bLg as well as for IFN. The results obtained with all the samples used in this study are summarized in Table 1. In the case of bLg, different samples prepared and dried from GuHCl containing solutions show different levels of shifts in the lmax that can be directly related to the extent of unfolding of the protein. In the case of IFN, the emission scan obtained from the PEG-precipitated and vacuumdried sample is entirely different when compared to IFN in solution or simply vacuum-dried IFN. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
898
SHARMA AND KALONIA
Table 1. Summary of the Trp Fluorescence Emission lmax Obtained for Various Solution and Dried Samples of bLg and IFN Along with Moisture Content of the Dried Samples Protein bLg
bLg and 1.5 M GuHCl
bLg and 3.0 M GuHCl
bLg and 6.0 M GuHCl
IFN PEG-precipitated IFN
Sample Physical State
Moisture,a % w/w
Solution, pH 7.4 Lyophilized Vacuum dried, no GuHCl Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution, pH 7.4 Vacuum dried, 10 h Vacuum dried, 24 h Solution Vacuum dried from solution Wet precipitate Vacuum dried Vacuum dried with mannitol Vacuum dried with trehalose
— 1.5 2.1 — 2.5 1.6 — 16.5 10 — 29.5 20 — 1.6 — 2.9 1.5 2.5
lmax, nm 336 334 332 335 343.5 336 347.5 343.5 345.5 350.5 345.5 345.5 336 332 336 332, 352b 332, 352b 332
a
All samples were done in triplicate and the standard deviation was less than 5% of the average. The two values for lmax correspond to the two peaks in the emission scan.
b
This is a clear indication of significant changes in the tertiary structure of protein and also establishes the application of fluorescence spectroscopy to monitor these changes in solid samples. The inability of mannitol to preserve IFN structure and the ability of trehalose to maintain protein structure during drying is further demonstrated in the fluorescence emission scans from these samples. It can be concluded that the use of Trp fluorescence spectroscopy on protein powders can provide important insight into the changes in the tertiary structure pf protein molecules in the solid state. Furthermore, the technique can be extended to proteins in a semisolid or partially dried state such as wet precipitates to probe protein tertiary structure. We hope that this technique provides enormous help to the protein formulation scientists to understand and maintain the long-term stability of protein therapeutics.
ACKNOWLEDGMENTS The authors gratefully acknowledge financial support from the National Science Foundation Industry/University Cooperative Research Center for Pharmaceutical Processing and Hoffman-La JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Roche for the donation of interferon alpha-2a samples.
REFERENCES 1. Manning MC, Patel K, Borchardt RT. 1989. Stability of protein pharmaceuticals. Pharm Res 6:903–918. 2. Arakawa T, Prestrelski S, Kenney WC, Carpenter JF. 1993. Factors affecting short-term and longterm stabilities of proteins. Adv Drug Del Rev 10:1–28. 3. Wang W. 1999. Instability, stabilization, and formulation of liquid protein pharmaceuticals. Int J Pharm 185:129–188. 4. Pikal MJ. 1994. Freeze-drying of proteins: Process, formulation, and stability. In: Cleland JL, Langer R, editors. Formulation and delivery of proteins and peptides. Washington, DC: American Chemical Society, pp 120–133. 5. Wang W. 2000. Lyophilization and development of protein pharmaceuticals. Int J Pharm 203:1–60. 6. Mumenthaler M, Hsu CC, Pearlman R. 1994. Feasibility study on spray-drying protein pharmaceuticals: Recombinant human growth hormone and tissue 0 type plasminogen activator. Pharm Res 11:12–20. 7. Maa YF, Nguyen P-A, Andya JD, Dasovich N, Sweeney TD, Shire SJ, Hsu CC. 1998. Effect of
STEADY-STATE TRYPTOPHAN FLUORESCENCE SPECTROSCOPY STUDY
8.
9.
10.
11.
12.
13.
14. 15.
16.
17.
18.
19.
spray drying and subsequent processing conditions on residual moisture content and physical/biochemical stability of protein inhalation powders. Pharm Res 15:768–775. Dong A, Huang P, Caughey WS. 1990. Protein secondary structures in water from second-derivative amide I infrared spectra. Biochemistry 29:3303– 3308. Prestrelski SJ, Tedeschi N, Arakawa T, Carpenter JF. 1993. Dehydration-induced conformational transitions in proteins and their inhibition by stabilizers. Biophys J 65(2):661–671. Prestrelski SJ, Arakawa T, Carpenter JF. 1993. Separation of freezing- and drying-induced denaturation of lyophilized proteins using stress specific stabilization: II. Structural studies using infrared spectroscopy. Arch Biochem Biophys 303:465– 473. Dong A, Caughey WS. 1994. Infrared methods for study of hemoglobin reactions and structures. Methods Enzymol 232:139–175. Griebenow K, Klibanov AM. 1995. Lyophilizationinduced reversible changes in the secondary structure of proteins. Proc Natl Acad Sci USA 92: 10969–10976. Allison SD, Chang B, Randolph TW, Carpenter JF. 1999. Hydrogen bonding between sugar and protein is responsible for inhibition of dehydrationinduced protein unfolding. Arch Biochem Biophys 365:289–298. Ptitsyn OB. 1995. Molten globule and protein folding. Adv Protein Chem 47:83–229. Fink AL. 1995. Compact intermediate states in protein folding. Annu Rev Biophys Biomol Struct 24:495–522. Smith SO, Griffin RG. 1988. High-resolution solidstate NMR of proteins. Annu Rev Phys Chem 39:511–535. Derome AE, Bowden S. 1991. Low-temperature solid-state NMR of proteins. Chem Rev 91:1307– 1320. Fu R, Cross TA. 1999. Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure. Annu Rev Biophys Biomol Struct 28: 235–268. Tycko R. 2001. Biomolecular solid state NMR: Advances in structural methodology and applications to peptide and protein fibrils. Annu Rev Phys Chem 52:575–606.
899
20. Strambini EG, Strambini GB. 2000. Tryptophan phosphorescence as a monitor of protein conformation in molecular films. Biosensors Bioelectron 15:483–490. 21. Gabellieri E, Strambini GB. 2001. Structural perturbations of azurin deposited on solid matrices as revealed by Trp phosphorescence. Biophys J 80:2431–2438. 22. Strambini GB, Gabellieri E. 1984. Intrinsic phosphorescence from proteins in the solid state. Photochem Photobiol 39:725–729. 23. Buijs J, Hlady V. 1997. Adsorption kinetics, conformation, and mobility of the growth hormone and lysozyme on solid surfaces, studied with TIRF. J Colloid Interface Sci 190:171–181. 24. Lakowicz JR. 1982. Principles of fluorescence spectroscopy. New York: Plenum Press. 25. Eftink MR. 1994. The use of fluorescence methods to monitor unfolding transitions in proteins. Biophys J 66:482–501. 26. Haria M, Benfield P. 1995. Interferon a-2a: A review of its pharmacological properties and therapeutic use in the management of viral hepatitis. Drugs 50:873–896. 27. Klaus W, Gsell B, Labhardt AM, Wipf B, Senn H. 1997. The three-dimensional high-resolution structure of human interferon a-2a determined by heteronuclear NMR spectroscopy in solution. J Mol Biol 274:661–675. 28. Uhrinova S, Smith MH, Jameson GB, Uhrin D, Sawyer L, Barlow PN. 2000. Structural changes accompanying pH-induced dissociation of the Blactoglobulin dimer. Biochemistry 39:3565–3574. 29. Tanford C, Kawaharo K, Lapanje S. 1966. Proteins in 6M guanidine hydrochloride. Demonstration of random coil behavior. J Biol Chem 241:1921–1923. 30. Tanford C, Kawaharo K, Lapanje S. 1967. Proteins as random coils. I. Intrinsic viscosities and sedimentation coefficients in concentrated guanidine hydrochloride. J Am Chem Soc 89:729–736. 31. Nozaki Y, Tanford C. 1967. Proteins as random coils. II. Hydrogen ion titration curve of ribonuclease in 6M guanidine hydrochloride. J Am Chem Soc 89:742–749. 32. Tanford C, Kawahara K, Lapanje S, Hooker TM, Zarlengo MH, Salahuddin A, Aune KC, Takagi T. 1967. Proteins as random coils. III. Optical rotatory dispersion in 6M guanidine hydrochloride. J Am Chem Soc 89:5023–5029.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003