Anal Bioanal Chem DOI 10.1007/s00216-017-0390-y
RESEARCH PAPER
Solid-film sampling method for the determination of protein secondary structure by Fourier transform infrared spectroscopy Junting Zhang 1 & Xiaoning Zhang 1 & Fan Zhang 2 & Shaoning Yu 1
Received: 2 March 2017 / Revised: 23 April 2017 / Accepted: 2 May 2017 # Springer-Verlag Berlin Heidelberg 2017
Abstract Fourier transform infrared (FTIR) spectroscopy is one of the widely used vibrational spectroscopic methods in protein structural analysis. The protein solution sample loaded in demountable CaF2 liquid cell presents a challenge and is limited to high concentrations. Some researchers attempted the simpler solid-film sampling method for the collection of protein FTIR spectra. In this study, the solid-film sampling FTIR method was studied in detail. The secondary structure components of some globular proteins were determined by this sampling method, and the results were consistent with those data determined by the traditional solution sampling FTIR method and X-ray crystallography, indicating that this sampling method is feasible and efficient for the structural characterization of proteins. Furthermore, much lower protein concentrations (~0.5 mg/mL) were needed to obtain highquality FTIR spectra, which expands the application of FTIR spectroscopy to almost the same concentration range used for circular dichroism and fluorescence spectroscopy, making comparisons among three commonly used techniques possible in protein studies.
Keywords FTIR spectroscopy . Protein structure . Protein secondary structural composition . Protein concentration Electronic supplementary material The online version of this article (doi:10.1007/s00216-017-0390-y) contains supplementary material, which is available to authorized users. * Shaoning Yu
[email protected] 1
Department of Chemistry, Fudan University, Shanghai 200433, China
2
Zhejiang BioHarmonious SciTech. Co. LTD., Hangzhou, Zhejiang 310018, China
Introduction Fourier transform infrared (FTIR) spectroscopy is one of the widely used vibrational spectroscopic methods in protein structural analysis [1–6]. Early experiments were limited to determine the secondary structure in solid states or in deuterated aqueous states [2, 7, 8], mainly due to strong water absorption overlaps in the protein amide I region (1600 to 1700 cm−1), which has been the most useful probe for determining the secondary structures of proteins in solution [9, 10]. For the study of proteins in aqueous solution by FTIR spectroscopy, the accuracy of subtracting large H2O bands maybe the greatest problem. Previous studies have suggested that the contradiction can be avoided by using D2O as a solvent because there is no absorptional spectrum for D2O in the region where the amide I and II bands are observed [8, 11]. However, this would alter protein secondary structures because of the impact of the incomplete exchange of D for H on the strength and length of hydrogen bonds. Therefore, H2O-based media have the advantage of providing a more native environment. The greater sensitivity of the FTIR instrument and new H2O subtraction program make it possible to obtain a highquality protein FTIR spectrum in aqueous solution [11–13]. The water absorbance can be quantitatively subtracted using an IR cell of sufficiently small path length (6–10 μm). Using the demountable CaF2 liquid cell including two rectangular windows and an appropriate spacer (6–10 μm), the background can be subtracted accurately from the protein spectrum using strict procedures [12–15]. Despite the extensive applications of the cell, a few studies have suggested that it still has practical problems [16–20]. The short path length makes it more difficult to match the path lengths between sample and reference cells [19, 20]. In addition, this method makes it difficult to fill the entire window surface without air bubbles or unfilled spaces [21]. On the other hand, the protein
Zhang J. et al.
Protein purity was assessed by staining SDS-PAGE gels with Coomassie Blue. Purified protein was stored at −80 °C before use. The additives in protein solutions were removed by dialysis against a buffer containing 50 mM Tris (pH 7.5) and 100 mM NaCl. The protein was concentrated to ~10 mg/mL by an Amicon Ultra-4 centrifugal filter unit (MWCO 3 kDA, Millipore). The concentrations were determined spectrophotometrically. The protein powders were dissolved in 50 mM Tris (pH 7.5) and 100 mM NaCl, and the final protein concentration was in the range of 0.5–20 mg/mL.
0.08
A 0.06 0.04 0.02 0.00 1800
1700
1600
1500
1400
1300
B
-d A/dv
Sample preparation
The final concentration of protein in samples prepared for infrared measurement ranged from 0.5 to 10 mg/mL. FTIR spectra were recorded at 20 °C using a Bomem MB series FTIR Spectrometer (Quebec, Canada) equipped with a dTGS detector and purged constantly with dry air. For each spectrum, a 128-scan interferogram was collected in singlebeam mode with 4 cm−1 resolution. Reference spectra were recorded under identical conditions with only the corresponding buffer in the cell. Protein spectra were obtained using a previously established protocol [30, 31]. For the sake of comparison, the spectra were collected under the same conditions for both the solid-film sampling method and traditional H2Obased aqueous sampling method.
Absorbance
Lysozyme (chicken egg white, 62,970), myoglobin (equine skeletal muscle, M0630), hemoglobin (horse, H4632), trypsin (bovine, T1426), BSA (bovine, A7030), HSA (human, A9731), IgG (human, 56,834), and cytochrome c (horse heart, C7752) were purchased from Sigma and used without further purification. Recombinant CRP was expressed and purified as described previously [28]. All other chemicals were reagent grade and of the highest purity commercially available.
FTIR spectroscopy
2
Materials and methods
mg/mL) was applied to the surface of the rectangular CaF2 plate and the liquid spread with a tip. The area of the sample was approximately 4 × 4 mm2 in the center of the CaF2 plate. The free H2O in the sample was evaporated naturally at room temperature. During this process, the bound H2O was retained in the sample due to hydrogen bonds between the protein and H2O [29]. After drying (1–3 h), a thin solid film was formed on the surface of the CaF2 plate, which was then mounted in a cell holder for IR spectrum collection. Buffer sample was applied in the same way as a reference.
2
concentration should be high enough (>5 mg/mL) in aqueous solution [22]. The concentration requirement renders the technique unsuitable for studies of the secondary structures of proteins in aqueous solution at low concentrations [23–25]. Some researchers have attempted the simpler solid-film sampling method for the collection of protein FTIR spectra [26, 27]. Due to the complexity of the traditional solution sampling method, the solid-film sampling maybe an alternative method for determining protein secondary structure by FTIR spectroscopy. The goal of this work was to systematically investigate the feasibility of this simpler sampling method. In this study, the secondary structure components of some globular proteins, namely myoglobin, hemoglobin, human serum albumin (HSA), immunoglobulin G (IgG), recombinant cAMP receptor protein (CRP), bovine serum albumin (BSA), lysozyme, trypsin, chymotrypsin, and cytochrome c, were measured by the solid-film sampling method. The effect of evaporation time for solid-film formation and the protein concentration range were also addressed in this work.
Solid-film method The traditional cell used for IR spectroscopy is a demountable transmission CaF 2 cell and the procedure followed the established protocol [21]. For solid-film sampling, a single CaF2 plate was cleaned with double distilled water and ethanol. After drying, 10–20 μl of the protein solution (0.5–10
1690 1680 1670 1660 1650 1640 1630 1620 1610 -1
Wavenumber(cm )
Fig. 1 FTIR absorbance (A) and inverse second-derivative (B) spectra of myoglobin (10 mg/mL, pH 7.5) determined by the solid-film sampling method
Solid-film sampling method for the determination of protein
Hemoglobin
HSA
Lysozyme
Trypsin
Cytochrome c
CRP
Chymotrypsin
IgG
-d A/dv
BSA
1690
1664
1638
1612
1690
1664
1638
1690
1612
1664
1638
1612
-1
Wavenumber(cm ) Fig. 2 Comparisons of the inverse second-derivative spectra of nine selected proteins in aqueous solution using the solid-film sampling (solid lines) and traditional solution sampling (dashed lines). The protein concentration is ~8 mg/mL, pH 7.5
Subtraction of the reference spectrum from the protein solution spectrum was carried out in accordance with the criteria described previously. The resulting protein difference spectra were smoothed with a nine-point Savitsky-Golay function to remove potential white noise [32]. Second-derivative spectra Table 1
were obtained with a seven-point Savitsky-Golay derivative function, baseline-corrected and area-normalized as described previously [33]. The relative secondary structure content was determined from a curve-fitting analysis of the inverted second-derivative spectrum in the amide I band range of
Distributions of secondary structures of selected proteins determined by two FTIR sampling methods and X-ray crystallography
Protein
Hemoglobin Lysozyme HSA Trypsin IgG BSA Chymotrypsin CRP Cytochrome C
α-Helix
β-Sheet
β-Turn
Random
X-ray
FTIRa
FTIRb
X-ray
FTIRa
FTIRb
X-ray
FTIRa
FTIRb
X-ray
FTIRa
FTIRb
87 45 69 9 3 87 8 37 48
79 60 80 20 3 85 7 38 54
78 59 86 21 3 86 3 35 42
0 19 0 56 67 0 50 38 10
7 14 7 51 76 5 58 41 16
12 12 5 44 64 6 51 42 21
7 23 19 24 18 7 27 19 17
14 7 14 23 21 11 23 15 18
10 10 9 20 28 9 28 16 25
6 13 12 11 12 8 15 6 25
0 19 0 6 0 0 12 4 12
0 19 0 15 5 0 17 6 12
X-ray data are from Levitt and Greer [34] a
Data obtained from solid-film sampling
b
Data obtained from traditional solution sampling
Absorbance
Zhang J. et al.
0.2
A
was fair enough and the water vapor quantitatively subtracted. Quantitative analysis of the new method revealed that myoglobin contained approximately 79% α-helices and 21% β structures (β sheet and β turn). This result is consistent with the data from X-ray crystallography [30].
1h 5h 10h 15h 24h
0.1
The secondary structure components of some globular proteins determined by solid-film sampling method
0.0 1800
1700
1600
1500
1400
B
1300
-d2A/dv2
1h 5h 10h 15h 24h
1680
1660
1640
1620 -1
Wavenumber(cm ) Fig. 3 FTIR absorbance (A) and inverse second-derivative (B) spectra of hemoglobin (10 mg/mL) acquired at different evaporation times
1600 to 1700 cm−1, which is predominantly ascribed to >C = O stretching vibration of the peptide bond. The band area for each component peak was used to calculate the relative contribution of components to a particular protein secondary structure. BOMEM GRAMS/32 software (ABB Bomen, Inc.) was used for data acquisition and analysis.
To further investigate whether the new sampling method works, the FTIR spectra of hemoglobin, HSA, IgG, recombinant CRP, BSA, lysozyme, trypsin, chymotrypsin, and cytochrome c were recorded using the solid-film sampling method. Hemoglobin, HSA, and BSA are primarily α-helical structure proteins. IgG is classified as a β-structure protein. Lysozyme, cytochrome c, a-chymotrypsin, CRP, and trypsin have been classified as α + β proteins. Their three-dimensional structures are known from the PDB database [34, 35]. The original FTIR absorbance spectra of nine selected proteins are shown in Fig. S1 (see Electronic Supplementary Material, ESM). The second-derivative FTIR spectra of these proteins are shown in Fig. 2 and compared to the spectra determined by the traditional solution sampling. The relative amounts of secondary structural component in these proteins as determined by different FTIR sampling methods and calculated from crystallography data are listed in Table 1 [21]. Taking the experimental error into account, the results from the different FTIR sampling methods and X-ray crystallography methods are in good agreement. Thus, the simpler solid-film sampling method is fair enough to estimate the secondary structural components of proteins. The solid-film sampling method was successfully used in the attenuated total reflectance FTIR (ATR-FTIR) technique
Results and discussion
The distribution of secondary structures determined from the amide I spectra of globular proteins provides support for the utility of the reported infrared second-derivative spectra (IRSD) method for the qualitative and quantitative analysis of protein secondary structures in solution [31]. Myoglobin is monomeric molecules that contain a single-chain polypeptide, which has been known for decades from FTIR spectroscopy and X-ray crystallographic studies. In this assay, myoglobin was used to test the feasibility of the sampling method. Approximately 20 μl of the myoglobin solution (10 mg/ mL) and buffer was applied to the surface of a rectangular CaF2 plate as described above. After 1 h, the protein and buffer solution formed a shining solid film. The FTIR spectrum was flat and smooth (Fig. 1) and the baseline between 2000 and 1750 cm−1 straight, indicating that the protein signal
Absorbance 1800
0.3mg/ml 0.5mg/ml 1mg/ml 3mg/ml 5mg/ml 10mg/ml
-d2A/dv2
FTIR spectra of proteins determined by solid-film sampling method
1680
1660
164 0
1620
Wavenumber(cm-1 )
1700
1600
1500
1400
1300
Wavenumber(cm-1) Fig. 4 The absorbance spectra of BSA acquired at various concentrations using solid-film sampling method. The inset shows the inverse second-derivative spectra of BSA in the amide I region. The amide I absorbance maxima of all spectra were normalized to the same intensity for ease of comparison
Solid-film sampling method for the determination of protein
the ESM. There was no visible H2O in the hemoglobin sample after 1 h of evaporating the 10 mg/mL protein solution. No differences in FTIR absorbance and second-derivative spectra were observed (Fig. 3) when recording in 24 h. After natural evaporation, the sample can still be considered hydrated protein with bound H2O. The distribution of secondary structure components was the same (87% α-helices and 13% βstructures, data not shown). Thus, the evaporation time will not change the FTIR spectrum within 24 h. In order to prevent loss of bound H2O, the spectrum should be collected within 24 h.
Absorbance
-d 2A/dv 2
0.5mg/ml 1mg/ml 3mg/ml 5mg/ml 10mg/ml
1680
1660
1640
1620
Wavenumber(cm-1)
1800
1700
1600
1500
1400
1300
Effect of protein concentration on FTIR spectra
Wavenumber(cm-1)
High protein concentrations are considered a shortcoming of the FTIR technique to estimate the secondary structure of proteins [38]. As the protein concentration decreases, the quality of the FTIR spectrum deteriorated due to a decreasing signal/noise ratio, especially in the protein conformationsensitive amide I region, which would affect the accuracy of data interpretation [38]. In order to investigate the concentration limitation of the solid-film FTIR sampling method, the FTIR absorbance and second-derivative spectra were determined for BSA at 0.3, 0.5, 1, 3, 5, and 10 mg/mL (Fig. 4). When the protein concentration was >0.5 mg/mL, the quality of the absorbance and second-derivative spectra was feasible. However, the quality of the FTIR spectrum deteriorated when the protein concentration was 0.3 mg/mL. Therefore, when using the solid-film sampling method, FTIR spectra should be collected with protein concentrations >0.5 mg/mL. The FTIR spectra of different concentrations of BSA determined by the traditional solution sampling method are shown in Fig. 5 as a comparison. The distributions of secondary structure components estimated from the two sampling methods are listed in Table 2. The secondary structure components of BSA were consistent with the data obtained using the traditional solution sampling when the concentration was >5 mg/mL. The solid film highly concentrated the proteins, with a high signal-to-noise ratio of protein/buffer, making it possible to obtain high-quality spectra at low original protein
Fig. 5 The absorbance and inverse second-derivative spectra of BSA acquired at various concentrations using the traditional solution sampling method. The inset shows the second-derivative spectra of BSA in the amide I region. The amide I absorbance maxima of all spectra were normalized to the same intensity for ease of comparison
for determining the secondary structure of proteins in inclusion bodies, protein folding aggregates, and amyloid fibrils. ATR-FTIR samples containing materials with a high refractive index may have significant shifts in the observed band positions, making it difficult to make comparisons between samples [36, 37]. In the transmission FTIR method, the results from solid-film sampling were consistent with the data determined by the traditional solution sampling FTIR method and X-ray crystallography, indicating that the solid-film formation did not change the secondary structures of selected proteins.
Effect of evaporation time on FTIR spectra For solid-film sampling method, the evaporation time depended on the protein properties and concentration. Generally, the FTIR spectrum could be recorded when there was no visible H2O in the sample. In order to investigate the effect of evaporation time on FTIR spectra, spectra were recorded for hemoglobin at different evaporation times (Fig. 3). The full-range spectrum of hemoglobin is shown in Fig. S2 in Table 2 Secondary structure content of BSA solutions at different concentrations determined by FTIR spectroscopy with two sampling methods
BSA concentration (mg/mL)
0.5 1 3 5 10
α-Helix
β-Sheet
β-Turn
Random
FTIRa
FTIRb
FTIRa
FTIRb
FTIRa
FTIRb
FTIRa
FTIRb
83 87 82 89 88
25 58 63 80 87
9 7 10 6 6
54 34 33 13 6
8 6 8 5 6
17 8 4 7 7
0 0 0 0 0
8 0 0 0 0
a
Data obtained from solid-film sampling
b
Data obtained from traditional solution sampling
Zhang J. et al.
concentration with this method. This method is particularly suitable for proteins with low solubility. Thus, the solid-film sampling method greatly expanded the concentration range for application of FTIR spectroscopy in protein structure studies, especially for proteins with low solubility. Furthermore, in this concentration range, the FTIR results are comparative with the results of circular dichroism and fluorescence spectroscopy.
5.
6.
7.
8. 9.
Conclusions In summary, the solid-film method has been studied systematically and the process of preparing sample was described in detail. The relative amounts of secondary structure components in select globular proteins determined by the solid-film sampling method agree closely with the data determined by the reported FTIR solution injection sampling method and Xray crystallography. Taken together, the results suggest that the solid-film FTIR sampling method is feasible for the structural characterization of proteins. This sampling method can be used with soluble globular proteins, and the process of sample preparation did not alter the secondary structures of these selected proteins. Moreover, the solid-film method renders the FTIR technique suitable for determining the secondary structure of proteins in aqueous solutions at lower concentrations (>0.5 mg/mL), much lower than the traditional method. It expands the application of FTIR spectroscopy to almost the same concentration range used for circular dichroism and fluorescence spectroscopy, making comparisons among three commonly used techniques possible in protein studies.
10. 11.
12.
13.
14.
15.
16.
17.
18. Acknowledgements This work was supported by the National Natural Science Foundation of China (Nos. 31470786 and 21275032).
19.
Compliance with ethical standards 20. Conflict of interest The authors declare that they have no conflict of interest.
References 1. 2.
3.
4.
Smith BC. Fundamentals of Fourier transform infrared spectroscopy. Boca Raton: CRC Press; 1996. Surewicz WK, Mantsch HH. New insight into protein secondary structure from resolution-enhanced infrared spectra. Biochim Biophys Acta Protein Struct Mol Enzymol. 1988;952:115–30. Zhao Y, Yang H, Meng K, Yu S. Probing the Ca 2+/CaM-induced secondary structural and conformational changes in calcineurin. Int J Biol Macromol. 2014;64:453–7. Fu C, Zhang J, Zheng Y, Xu H, Yu S. Binding of calmodulin changes the calcineurin regulatory region to a less dynamic conformation. Int J Biol Macromol. 2015;79:235–9.
21.
22.
23.
24.
25.
Jiang Y, Li C, Nguyen X, Muzammil S, Towers E, Gabrielson J, et al. Qualification of FTIR spectroscopic method for protein secondary structural analysis. J Pharm Sci. 2011;100(11):4631–41. Goormaghtigh E, Ruysschaert J-M, Raussens V. Evaluation of the information content in infrared spectra for protein secondary structure determination. Biophys J. 2006;90(8):2946–57. Parker FS. Applications of infrared, Raman, and resonance Raman spectroscopy in biochemistry. Springer Science & Business Media; 1983. Byler DM, Susi H. Examination of the secondary structure of proteins by deconvolved FTIR spectra. Biopolymers. 1986;25(3):469–87. Susi H. Infrared spectroscopy—conformation. Methods Enzymol. 1972;26:455–72. Griffiths P, Durig J. Analytical applications of FT-IR to molecular and biological systems. Dordrecht: Reidel; 1980. 11 Susi H, Timasheff SN, Stevens L. Infrared spectra and protein conformations in aqueous solutions I. The amide I band in H2O and D2O solutions. J Biol Chem. 1967;242(23):5460–6. Gorga JC, Dong A, Manning MC, Woody RW, Caughey WS, Strominger JL. Comparison of the secondary structures of human class I and class II major histocompatibility complex antigens by Fourier transform infrared and circular dichroism spectroscopy. Proc Natl Acad Sci India. 1989;86(7):2321–5. Lee DC, Haris PI, Chapman D, Mitchell RC. Determination of protein secondary structure using factor analysis of infrared spectra. Biochemistry. 1990;29(39):9185–93. Yamamoto T, Tasumi M. FT-IR studies on thermal denaturation processes of ribonucleases A and S in H2O and D2O solutions. J Mol Struct. 1991;242:235–44. Kong J, Yu S. Fourier transform infrared spectroscopic analysis of protein secondary structures. Acta Biochim Biophys Sin. 2007;39(8):549–59. Fu C, Yang H, Wang M, Xiong H, Yu S. Serum albumin adsorbed on Au nanoparticles: structural changes over time induced by S–Au interaction. Chem Commun. 2015;51(17):3634–6. Wang M, Fu C, Liu X, Lin Z, Yang N, Yu S. Probing the mechanism of plasma protein adsorption on Au and Ag nanoparticles with FT-IR spectroscopy. Nano. 2015;7(37):15191–6. Wu G, Gao Z, Dong A, Yu S. Calcium-induced changes in calmodulin structural dynamics and thermodynamics. Int J Biol Macromol. 2012;50(4):1011–7. Dong A, Huang P, Caughey WS. Redox-dependent changes in. beta.-extended chain and turn structures of cytochrome c in water solution determined by second derivative amide I infrared spectra. Biochemistry. 1992;31(1):182–9. Kal'Nin N, Ven'yaminov SY. Quantitative measurement of the IR spectra of water solutions. J Appl Chem. 1988;49(4):1028–32. Yang H, Yang S, Kong J, Dong A, Yu S. Obtaining information about protein secondary structures in aqueous solution using Fourier transform IR spectroscopy. Nat Protoc. 2015;10(3):382–96. Dong A, Randolph TW, Carpenter JF. Entrapping intermediates of thermal aggregation in alpha-helical proteins with low concentration of guanidine hydrochloride. J Biol Chem. 2000;275(36):27689–93. Ismail AA, Mantsch HH, Wong PT. Aggregation of chymotrypsinogen: portrait by infrared spectroscopy. Biochim Biophys Acta Protein Struct Mol Enzymol. 1992;1121(1):183–8. Fabian H, Mantsch HH. Ribonuclease A revisited: infrared spectroscopic evidence for lack of native-like secondary structures in the thermally denatured state. Biochemistry. 1995;34(41):13651–5. Martínez A, Haavik J, Flatmark T, Arrondo JLR, Muga A. Conformational properties and stability of tyrosine hydroxylase studied by infrared spectroscopy EFFECT OF IRON/CATECHOLAMINE BINDING AND PHOSPHORYLATION. J Biol Chem. 1996;271(33):19737–42.
Solid-film sampling method for the determination of protein 26.
27.
28.
29.
30.
31.
Lee TH, Cheng WT, Lin SY. Thermal stability and conformational structure of salmon calcitonin in the solid and liquid states. Biopolymers. 2010;93(2):200–7. Seshadri S, Khurana R, Fink AL. Fourier transform infrared spectroscopy in analysis of protein deposits. Methods Enzymol. 1999;309:559–76. Aiba H, Fujimoto S, Ozaki N. Molecular cloning and nucleotide sequencing of the gene for E. coli cAMP receptor protein. Nucleic Acids Res. 1982;10(4):1345–61. Luthra S, Kalonia DS, Pikal MJ. Effect of hydration on the secondary structure of lyophilized proteins as measured by Fourier transform infrared (FTIR) spectroscopy. J Pharm Sci. 2007;96(11): 2910–21. Dong A, Caughey WS. Infrared methods for study of hemoglobin reactions and structures. Methods Enzymol. 1994;232: 139–75. Dong A, Huang P, Caughey WS. Protein secondary structures in water from second-derivative amide I infrared spectra. Biochemistry. 1990;29(13):3303–8.
32.
Savitzky A, Golay MJ. Smoothing and differentiation of data by simplified least squares procedures. Anal Chem. 1964;36(8):1627–39. 33. Dong A, Prestrelski SJ, Allison SD, Carpenter JF. Infrared spectroscopic studies of lyophilization-and temperature-induced protein aggregation. J Pharm Sci. 1995;84(4):415–24. 34. Levitt M, Greer J. Automatic identification of secondary structure in globular proteins. J Mol Biol. 1977;114(2):181–239. 35. Richardson JS. The anatomy and taxonomy of protein structure. Adv Protein Chem. 1981;34:167–339. 36. Jackson M, Mantsch HH. The use and misuse of FTIR spectroscopy in the determination of protein structure. Crit Rev Biochem Mol. 1995;30(2):95–120. 37. Uversky VN, Permiakov EAe. Methods in protein structure and stability analysis: luminescence spectroscopy and circular dichroism. Nova Publishers; 2007. 38. Dong A, Jones LS, Kerwin BA, Krishnan S, Carpenter JF. Secondary structures of proteins adsorbed onto aluminum hydroxide: infrared spectroscopic analysis of proteins from low solution concentrations. Anal Biochem. 2006;351(2):282–9.
