488 Biochimica et Biophysica Acta 870 (1986) 488-494 Elsevier
Characterization and conformational analysis by Raman spectroscopy of human airway lysozyme
J. Marx a, j. Jacquot b, M. Berjot a, E. Puchelle b and A.J.P. Alix a a Groupe de Spectroscopie Vibrationnelle des Systbmes Bio-Molbculaires, GSV-SBM, Laboratoire de Recherches Optiques,
UER Sciences, BP 347, 51062 Reims Cedex, and b Groupe d'Etude des Systbmes Mucociliaires, GEMUC, Laboratoire d'Histologie, UER Mbdecine, 51095 Reims Cedex (France)
(Received July 26th, 1985) (Revised manuscript received December 6th, 1985)
Key words: Lysozyme secondary structure; Raman spectroscopy; Conformational analysis; Human airway; (Hen egg-white)
Human airway iysozyme, purified from pathological bronchial secretions, is characterized by a specific activity 3-fold higher than that of hen egg-white lysozyme. The amino acid composition of human airway lysozyme is identical to that of other human lysozymes. The laser Raman spectra of human airway lysozyme and hen egg-white lysozyme in phosphate buffer solution (pH 7.2) are recorded in the range 300-1900 cm-t at 488 nm. Drastic intensity differences are observed between the spectra analyzed in the ranges characteris- tic of the peptide backbone (e.g., l-sheet; Ca-C, Ca-N), and of the aromatic side-chain vibrations (tyrosine, tryptophan). The deconvolution of the Raman amide I band gives secondary structures of 38% and 39% a-helix, 25% and 20% fl-sheet, and 37% and 41% undefined structure for the human and hen lysozymes, respectively.
Although human lysozyme has been isolated, purified and characterized from various sources such as tears, saliva, milk and urine [1-3], human airway lysozyme has received little attention, ex- cept for its immunocytochemical localization and quantitation in pathological conditions [4-5]. Hu- man airway secretions contain a variety of locally produced proteins, including secretory IgA, lacto- ferrin and human airway lysozyme, which are potentially important in the lung microbial de- fense. Human airway lysozyme is an antibacterial enzyme secreted in large quantities (10-20 mg of lysozyme cleared from airways daily) essentially by submucosal tracheal glands, surface epithelium and pulmonary alveolar macrophages . Investi- gations on the conformational and three-dimen- sional structures of lysozyme have been carried
out mostly with enzyme of non-human origin (es- sentially from hen egg-white) by circular dichro- ism , nuclear magnetic resonance , X-ray crystallographic analysis  and Raman spec- troscopy . Only human lysozymes purified from urine of leukemic patients have been studied in crystallized forms by the X-ray diffraction tech- nique [11-12]. The usefulness of Raman spec- troscopy applied to the investigation of biological molecules lies in its capability to study the material in any phase (solid state, crystal; powder, lyophi- lized; aqueous solution) and under various experi- mental conditions (pH, temperature, concentra- tion). The X-ray technique is only applicable to crystallized substances, although practically all proteins in biological systems are in aqueous solu- tion. In addition, the question always arises of whether the conformational structure of biological proteins is really the same in crystals and in the
0167-4838/86/$03.50 1986 Elsevier Science Publishers B.V. (Biomedical Division)
natural state. In contrast, Raman spectroscopy can be applied to the conformational analysis of proteins in aqueous solution and, moreover, gives information on free and/or bound-molecules (molecular interactions).
Therefore, this technique is a powerful tool for studying the conformational structures (sec- ondary, tertiary) of a protein, as it provides infor- mation on the peptide backbone structure, the geometries of the disulfide S-S and/or the C-S bonds and the environment of some side chains such as those of tyrosine and tryptophan .
The purpose of the present study is first to characterize the bacteriolytic activity and molecu- lar properties of human airway lysozyme, and second to analyze the Raman spectrum of the lysozyme in solution, within the same experimen- tal conditions. In addition, the human airway lysozyme Raman spectrum is compared to that of hen egg-white lysozyme and the secondary struc- tures of both lysozymes are determined from the deconvolution of the Raman amide I band (Refer- ence Intensity Profiles Method ).
Materials and Methods
Preparation of lysozymes. Human airway lyso- zyme was prepared from a pool of 2 liters of purulent bronchial secretions collected in patients with chronic bronchitis. The purification proce- dure involved NaC1 extraction, trichloroaceric acid treatment and dialysis followed by cation-ex- change chromatography on CM-Sephadex C-25 resin and gel filtration chromatography on Sep- hadex G-75 Superfine gel (Pharmacia Fine Chem- icals, Uppsala), as previously described . The lysozyme was pure as judged by SDS-polyacryl- amide gel electrophoresis by the method of King and Laemmli . An apparent molecular weight of 15000 was estimated. The lytic activity of the lysozyme was evaluated by measuring the initial rate of lysis of an Micrococcus luteus cell wall suspension (Worthington Biochemicals Corp., Freehold) used as substrate according to a mod- ified procedure of Shugar . A 0.1 ml sample of test enzyme solution was added to 2.9 ml substrate (0.15 mg/ml) in 0.07 M sodium phosphate buffer (pH 7.0). Changes in absorbance (AA) at 450 nm were followed spectrophotometrically for 3 rain at
25C, and the initial velocity was measured to give AA/min. The specific activity was defined as units of enzyme activity per mg protein determined by the method of Lowry et al. , using hen egg- white lysozyme as standard. The amino acid analyses were carried out with a Biotronik LC 6000 analyser after hydrolysis of human airway lysozyme samples for 18, 48 and 96 h in 5.6 M HC1 under vacuum at 110C. Cysteine was de- termined as cysteic acid after performic acid oxidation. Tryptophan was determined spectro- photometrically according to Simpson et al. .
Twice-crystallized hen egg-white lysozyme was purchased from Worthington and was used without further purification.
Raman spectroscopy. Just before the experi- ment, the enzymes (human and hen) were dis- solved in 0.06 M sodium phosphate buffer, 0.15 M NaC1 (pH 7.20) at a concentration of 10 mg/ml and then centrifuged at 10000 g for 10 min to clarify the lysozyme solution. Raman spectra were recorded on a Coderg PHO spectrometer in the wavelength region 300-1900 cm-1 (HELLMA cell, 10 10 ram). A Coherent Radiation Model 52B Ar + laser at 600 mW power with the 488 nm line was used. In these conditions, no sample degrada- tion was detected by changes in specific lyric activity of the human or hen lysozymes or modifi-, cations in the ultraviolet absorption of either lysozyme, before or after irradiation. The signal pulse obtained by a cool photomultiplier EMI 9558 QB was amplified, counted digitally and then stored in the computer system (ALCYANE, MBC, France), which coordinates the scanning (steps of 2 cm-1 with maximum counting time of 2 s per step) and the spectral data acquisition. Several types of operations may be carried out on the stored spectra . Data accumulation, time aver- aging (16 scans), was used to obtain the spectra and to improve the signal-to-noise ratio. The spec- tra are recorded at room temperature with slit widths of 6 cm-1 or 4 cm-1. The Raman frequen- cies reported here are accurate to + 2 cm-1. The Raman spectral intensities were measured at a particular frequency by their heights above the adjusted baseline. For a broad band, such as the amide I band, the integrated intensity was ob- tained by calculating the band area.
Amide I band deconoolution. The secondary
structure of the enzymes was obtained by applying an original method, the 'Reference Intensity Pro- files Method' . The methods uses the Raman spectral intensities of the backbone amide I vibra- tions of pure classes of secondary structures (a- helix, fl-sheet, undefined conformations) as refer- ence intensity profiles. The reference intensity profiles were determined from the analysis of the Raman amide I bands of a basic set of proteins and polypeptides, the secondary structures of which were estimated from X-ray data. The struc- ture criteria used were identical to those defined by Levitt and Greer , except that all 'aperiodi- cal' conformations are put together in a single class called 'undefined' conformation. Using such reference intensity profiles for the quantitative estimation of the secondary structure gave correla- tion coefficients (X-ray versus Raman data) higher than 0.97 for each class of structure. The reference intensity profiles method may be of use to detect characteristic conformational changes when the calculated shifts are larger than 4%. The percent contents are given accurate to ___ 2%. The reference profiles have maximum intensities at 1650, 1673 and 1660 cm -1 for the a-helical, fl-sheet and the undefined conformation, respectively. The 'treated' experimental Raman amide I band of a protein arises from a linear combination of the reference intensities and thus the fitted coeffi- cients give the corresponding amounts of the dif- ferent types of secondary structure. In the case of the study of a protein in solution and in order to take into account the 'solvent' contribution (essen- tially due to the water symmetric bending vibra- tion at 1640 cm-1), the problem of finding a subtraction coefficient working on the overall spectrum is solved together with the secondary structure determination. One introduces the spec- trum of the solvent, in the range 1630-1700 cm -1 (normalized and with an adjusted baseline: zero- point intensity at 1840 cm -~) as a new external contribution to the amide I band region. The fitted coefficients of that contribution give the subtraction coefficient permitting us to obtain the 'pure' spectrum of the protein. The operations ('manual' and/or automated) are summarized as follows: (i) storage of the overall spectra (300-1900 cm-1) of the protein in solution and of the solvent; (ii) treatment of the spectrum of the protein in
solution: fluorescence background correction, baseline correction related to the zero intensity points following from the amino acid composi- tion; (iii) normalization of the amide I band (pro- tein in solution) in the range 1630-1700 cm -1 after subtraction of the residue contributions (vibrations of aromatic amino acid side-chains, e.g., Tyr, Trp, Phe); (iv) normalization of the solvent contribution in the range 1630-1700 cm-1; and (v) deconvolution of the amide I band (pro- tein + solvent), giving us the coefficients of sec- ondary structure of the protein and the subtrac- tion coefficient (related to the solvent).
Results and Discussion
Isolation and characterization of human airway lysozyme
The purification procedure is summarized in Table I. Only two chromatographic steps are re- quired for obtaining more than 240 mg of pure human lysozyme from 2 liters of purulent bronchial secretions. The purified lysozyme pre- paration is dialyzed against deionized water and lyophilized. A 1% solution of human airway lyso- zyme in 0.05 M sodium phosphate buffer (pH 6.25) had an absorbance of 22.4 at 280 nm which is slightly lower than that of 25.5, 25.6 and 26.4 obtained for human urine , milk [2,23,24] and hen egg-white lysozyme , respectively. The amino acid composition of human airway lyso- zyme (Table II) compared to that of hen egg-white lysozyme taken from the sequence reported by Canfield  shows greater differences in the number of aspartic acid (human, 16, and hen, 21) and tyrosine (human, 6, and hen, 3) residues. Human airway lysozyme has an amino acid com- position identical to that of other lysozymes of human origin which have been previously reported and isolated from milk and urine [2,22,24]. According to a recent review by Jollrs and Jollrs , all human lysozymes which have been studied, up to now, have the same amino acid sequence (primary structure), but differ in their amino acid sequence from hen egg-white lysozyme by 52 amino acid substitutions (about 40% of 130 posi- tions). From our data, the specific activity of human airway lysozyme is 3.1 as compared to the hen egg-white lysozyme reference value of 1.0. The
PURIFICATION OF HUMAN AIRWAY LYSOZYME
Step Volume Total Total Spec. Purification Recovery (ml) protein lysozyme activity a act. (fold) (%)
Dialysis extract 1350 8 343.0 931.5 0.1 1 100 After CM-Sephadex C-25 215 295.0 805.6 2.7 24.8 86.5 After Sephadex G-75 Superfine 86 247.0 770.1 3.1 28.3 82.7
a Total lysozyme activity expressed as hen egg-white lysozyme activity.
molecular weight and specific activity of human airway lysozyme agree with the values reported by Konstan et al.  for the lysozyme isolated from human tracheal explants.
Raman spectra The Raman spectra of human airway lysozyme
and of hen egg-white lysozyme are shown in the Fig. 1. Taking into account the composition and spectra of all amino acids, we calculated a spec- trum giving the contributions of the residues (by summing up their corresponding spectra) and thus we obtained a set of points of minimum intensity which permit us to adjust the baseline of both human and hen lysozyme spectra, Such a con- sistent procedure allows us to compare the intensi- ties and the shapes of the bands of the two enzymes, which show marked intensity differences in the ranges characteristic of the aromatic side- chains, the peptide backbone ((Ca-C), (Ca-N)) and the peptide bond vibrations (amide I, III).
The residues tyrosine, tryptophan and phenyl- alanine give prominent Raman lines at fixed fre- quencies which are readily identified. Some of the observed bands are sensitive to the intra-molecu- lar environment of the residues. The 836/856 doublet of the tyrosine residues was deconvolved for human airway lysozyme and hen egg-white lysozyme, respectively. The intensity ratio I(856)/I(836), (1.5 for human and 1.0 for hen) indicates that all the six tyrosine residues are exposed to solvent for human airway lysozyme, but only 2 out of 3 for hen egg-white lysozyme. X-ray diffraction studies  and ultraviolet dif- ference spectrophotometrically studies  have indicated that, for hen egg-white lysozyme, Tyr-20 and Tyr-23 are exposed and Tyr-53 is more buried in the molecule. The presence of the 1358 cm -1
line of tryptophan, associated with a strong line at 876 cm -1, gives indication of some buried tryp- tophan residues. Moreover, the intensity ratio I(876)/I(1450), (0.25 for human airway lysozyme and 0.50 for hen egg-white lysozyme) reveals that the tryptophan residues are more buried in the hen than in the human lysozyme. From X-ray and chemical modification studies , it appears that
AMINO ACID COMPOSITION OF HUMAN AIRWAY LYSOZYME
Number of amino acid residues/molecule
Human airway lysozyme
Amino acid Experimental Nearest residues values a integer
Hen egg-white lysozyme b
Asp 16.4 16 21 Thr 4.6 c 5 7 Ser 5.8 c 6 10 Glu 8.6 9 5
P ro 2.0 2 2 Gly 11.1 11 12 Ala 13.6 14 12 Cys 8.0 8 8 Val 8.8 d 9 6 Met 1.3 1-2 2 lie 4.7 d 5 6 Leu 7.7 8 8 Tyr 5.8 6 3 Phe 1.8 2 3 Lys 4.8 5 6 His 0.8 1 1 Arg 12.4 d 13-14 11 Trp 4.4 4-5 6
a Mean values of two separate determinations. b Values taken from the sequence in Ref. 26. c Extrapolated to zero hydrolysis. d 96 h hydrolysis.
only 1 tryptophan residue of hen egg-white lyso- zyme (Trp-28) is completely buried, and the others are more or less buried with the following decreas- ing order: Trp-108 = Trp-l l l > Trp-63 > Trp-123 > Trp-62. The ranges 900-1000 cm -1 (p(C~-C)) and 1050-1150 cm -1 (p(C~-N)) show drastic changes for the peptide backbone vibrations of human airway lysozyme as compared to hen egg- white lysozyme. This is due likely to a reorganiza- tion of the side-chain conformations. The similar- ity of the line at 934 cm-1 for the human and the hen lysozymes suggests that the a-helix content is similar in both enzymes. The differences in the intensities and the shapes of the bands 1330 and 1450 cm -1 for human airway lysozyme and hen egg-white lysozyme (respectively characteristic of the 8(CH) and 8(CH 2, CH3) vibrations) is a strong indication of two different organizations of the side chains. It is noticeable that the numbers of CH- and CH2-, CH3- groups are identical in both enzymes.
Amide I, I I I band analysis The amide III band of hen egg-white lysozyme
has a line at 1260 cm -1 shifted to the lower frequencies (1252 cm -1) for human airway lyso- zyme, indicating that the human lysozyme, has a more important contribution of fl-sheet structure than the hen lysozyme (Fig. 1).
The broad non-symmetrical amide I bands show a maximum at 1658 cm -1, and large differences in their integrated intensities. This remarkable fea- ture has already been pointed out in cases of major reorganizations of the peptide backbone and more precisely when the content of fl-struc- ture increases . The deconvolution of the amide I band by use of the 'Reference Intensity Profiles Method' (Fig. 2) gives 38% and 39% a-helix, 25% and 20% fl-sheet and 37% and 41% undefined structures of human airway lysozyme and hen egg-white lysozyme, respectively. Artymiuk and Blake  recently refined the structure of human urinary lysozyme at 1.5 ,~ resolution. Estimation
- - !N~
~,,;'~"' ~ l CM-I I I000 14C)0 1800
Fig. 1. Raman spectra of human airway lysozyme (unbroken line) and of hen egg-white lysozyme (dotted line) in the range 600-1800 cm-1 (excitation line 488 nm, power 600 roW, phosphate buffer solution, pH 7.2, 16 scans).
." ,.+ i I
, # i , ~ . x :.. / / \ k . . . - / . . . . - , . , . , \ . ....,
I , I I ~ l C M- I . , Is4o Is6o Isao 1700
Fig. 2. Deeonvolution of the amide I band of human airway lysozyme by the reference intensity profiles method (~, experimental; . . . . . . , calculated amide I bands; x x x, total a-helix; . . . . . , total fl-sheet; . . . . . . , undefined contri- butions).
of the percent contents of the secondary structure of human urinary lysozyme, from its Ramachadran '/', ~-plot , gives 38% a-helix and 27% fl-sheet. Percentages obtained for solid hen egg-white lysozyme and estimated from X-ray data were: 46% a-helix, 19% fl-sheet and 35% undefined structure . Recently, Williams  calculated from the amide I band of the hen lysozyme in solution percentages of 47% a-helix, 19% fl-sheet and 34% undefined structure. From our Raman data, the small but the same significant increase of fl-sheet structure for human airway lysozyme is observed in comparison with hen egg-white lyso- zyme. Gavilanes et al.  have recently reported a similar increase in fl-sheet structure for human urine lysozyme compared to hen egg-white lyso- zyme, by using the predictive methodology of Chou and Fasman .
In this study, we compared the enzymatic prop- erties and Raman spectrum of human airway lysozyme to that of hen egg-white lysozyme in order to investigate the structure-activity relation- ship of human and animal lysozymes. The changes in Raman lines due to the aromatic side-chain vibrations indicate that the microenvironments of aromatic groups are markedly different in the human and the hen lysozymes. These differences might be related to higher specific lytic activity of
human airway lysozyme compared to that of hen egg-white lysozyme.
The secondary structure estimated from our data for human airway lysozyme is in good agree- ment with the recent X-ray data . Thus, it seems reasonable to propose that human airway lysozyme and human urinary lysozyme are the same lysozyme molecule with identical primary and secondary structures.
Additional Raman spectroscopic studies after binding inhibitors or substrate to both human airway lysozyme and hen egg-white lysozyme might determine more precisely which tridimen- sional structure is responsible for the higher lytic activity of the human lysozyme.
This work was supported by grants (CRE No. 855002 and 855019 from INSERM). We wish to thank Jean-Made Toumier for fruitful discus- sions, Yves Meckler for his excellent technical assistance and A. Quiqueret for typing the manuscript.
1 Joll~s, J. and Joll~s, P. (1967) Biochemistry 6, 411-417 2 Parry, R.M., Chandan, R.C. and Shahani, K.M. (1969)
Arch. Biochem. Biophys. 103, 59-65 3 Gachon, A.M., Kpamegan, G. and Dastugue, B. (1983)
Clin. Chim. Acta 129, 189-195 4 Bowes, D. Clark, A.E. and Corrin, B. (1981) Thorax 36,
108-115 5 Jacquot, J., Tournier, J.M., Carmona, T.G., Puchelle, E.,
Chazalette, J.P. and Sadoul, P. (1983) Clin. Resp. Physiol. 19, 453-458
6 Konstan, M.W., Chert, P.W., Sherman, J.M., Thomassen, M.J., Wood, R.E. and Boat, T.F. (1981) Am. Rev. Respir. Dis. 123, 120-124.
7 Kato, S., Shimamoto, N. and Utiyama, H. (1984) Biochim. Biophys. Acta 784, 124-132.
8 Howarth, O.W. and Yun Lian, L. (1984) Biochemistry 23, 3522-3526
9 Blake, C.C.F., Koenig, D.F., Mair, G.A., North, A.C.T., Phillips, D.C. and Sarma, V.R. (1965) Nature 206, 757-761
10 Schmidt, H. and Bicker, L. (1979) Arch. Biochem. Biophys. 195, 205-210
11 Artymiuk, P.J., Blake, C.C.F., Grace, D.E.P., Oatley, S.J., Phillips, D.C. and Sternberg, M.J.E. (1979) Nature 280, 563-568
12 Artymiuk, P.J. and Blake, C.C.F. (1981) J. Moi. Biol. 152, 737-762
13 Tu, A.T. (1982) Raman Spectroscopy in Biology: Principles and applications. John Wiley and Sons, New York
14 Alix, A.J.P., Berjot, M. and Marx, J. (1985) in Spectroscopy of Biological Molecules (Alix, A.J.P., Bernard, L. and Manfait, M., eds.), pp. 149-154, John Wiley and Sons, New York
15 Jacquot, J., Tournier, J.M. and Puchelle, E. (1985) Infect. Immun. 47, 555-560
16 King, J. and Laemmli, U.K. (1971) J. Mol. Biol. 62, 465-473 17 Shugar, D. (1952) Biochim. Biophys. Acta 8, 302-309 18 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall,
R.J. (1951) J. Biol. Chem. 193, 265-275 19 Simpson, R.J., Neuberger, M.R. and Liu, T.Y. (1976) J.
Biol. Chem. 251, 1936-1940 20 Marx, J., Beaudoin, J.L., Merienne, E. and Hudry-Cler-
geon, G. (1978) Proceedings of the 6th International Con- ference on Raman Spectroscopy, pp. 102-103, Bangalore
21 Levitt, M. and Greer, J. (1977) J. Mol. Biol. 114, 181-203 22 Canfield, R.E., Collins, J.C. and Sobel, J.H. (1974) in
Lysozyme (Osserman, E.F., Canfield, R.E. and Beychok, S., eds.), pp. 63-70, Academic Press, New York
23 Joll~s, J. and Joll~s, P. (1971) Helv. Chim. Acta 54, 2668-2675
24 Joll~s, J. and Joll~s, P. (1972) FEBS Lett. 22, 31-33 25 Sophianopoulos, A.J., Rhodes, C.K., Hoicomb, D.N. and
Van Holde, K.E. (1962) J. Biol. Chem. 237, 1107-1112 26 Canfield, R.E. (1963) J. Biol. Chem. 238, 2698-2707 27 Joilrs, P. and Jollbs, J. (1984) Mol. Cell. Biochem. 63,
165-189 28 Imoto, T., Johnson, L.N., North, A.C.T., Phillips, D.C. and
Rupley, J,A. (1972) The Enzymes, 3rd Edn., Vol. 7 (Boyer, P.D., ed.), pp. 665-868 Academic Press, New York
29 Izumi, T. and Inoue, H. (1976) J. Biochem. 79, 1309-1321 30 Marx, J., Hudry-Clergeon, G., Capet-Antonini, F. and
Bernard, L. (1979) Biochem. Biophys. Acta 578, 107-115 31 Williams, R.W. (1983) J. Mol. Biol. 166, 581-603 32 Gavilanes, J.G., Arias, L.M. and Rodriguez, R. (1984)
Comp. Biochem. Physiol. 77B, 83-88 33 Chou, P.Y. and Fasman, G.D. (1978) Annu. Rev. Biochem.