- Email: [email protected]
Microchemical analysis of retina layers in pigmented and albino rats by Fourier transform infrared microspectroscopy
Biochimica et Biophysica Acta 1473 (1999) 409^417 www.elsevier.com/locate/bba
Microchemical analysis of retina layers in pigmented and albino rats by Fourier transform infrared microspectroscopy Steven M. LeVine a , Je¡rey D. Radel b , Joseph A. Sweat c , David L. Wetzel a
c;
*
Department of Molecular and Integrative Physiology and the Mental Retardation and Human Development Research Center, University of Kansas Medical Center, Kansas City, KS 66160, USA b Department of Occupational Therapy Education and the Mental Retardation and Human Development Research Center, University of Kansas Medical Center, Kansas City, KS 66160, USA c Microbeam Molecular Spectroscopy Laboratory, Shellenberger Hall, Kansas State University, Manhattan, KS 66506, USA Received 21 June 1999; received in revised form 30 September 1999; accepted 30 September 1999
Abstract Fourier transform infrared (FT-IR) microspectroscopy is a powerful technique that can be used to collect infrared spectra from microscopic regions of tissue sections. The infrared spectra are evaluated to chemically characterize the absorbing molecules. This technique can be applied to normal or diseased tissues. In the latter case, FT-IR microspectroscopy can reveal chemical changes that are associated with discrete regions of lesion sites, which can provide insights into the chemical mechanisms of disease processes. In the present study, FT-IR microspectroscopy was used to analyze sections of retina from normal (pigmented) and albino rats. The outer segments of retinas from pigmented animals were found to have unusually strong absorption values for CNC^H unsaturation and carbonyl functional groups. Docosahexaenoic acid (DHA), a major constituent of lipids in the outer segments, also had particularly high absorption values for these functional groups, which suggests that it is responsible for those enhanced absorption values. Absorbance values for the unsaturation and carbonyl functional groups were substantially reduced in the outer segments of retinas from albino animals. This finding, together with data from other studies on light-induced oxidative events in the retina, indicates a loss of DHA by a light-induced mechanism in albino animals. The outer nuclear layer had strong absorbance values for H^C^OH and PNO functional groups, which is likely due to the sugar phosphate backbone of DNA. The outer and inner plexiform layers were found to contain greater concentrations of CH2 and CNO functional groups than the outer and inner nuclear layers, which is due to the high concentration of synaptic connections in the former layers. In summary, FT-IR microspectroscopy revealed a unique chemical profile in the outer segments compared to other retinal layers, and this profile was altered in albino animals. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Retina; Docosahexaenoic acid; FT-IR microspectroscopy; Outer segment
1. Introduction Infrared spectroscopy is used to identify organic
* Corresponding author. Fax: +1 (785) 532-7010; E-mail: [email protected]
functional groups within a specimen. Fourier transform infrared (FT-IR) microspectroscopy, which combines infrared spectroscopy with microscopy and computer science, allows for chemical analyses to be performed on microscopic regions of samples [1,2]. This technique can be used on tissue sections, and the chemical composition of microscopic regions
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 9 ) 0 0 1 9 5 - 6
BBAGEN 24943 1-12-99
410
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
can be compared to histological structures [1^3]. This chemical information is obtained in the absence of stains or ¢xatives. In addition to probing individual microscopic structures, experiments can be performed that collect spectra along a grid pattern across an area of interest in a tissue section. Interpolation of data between points is used to generate maps that reveal the distribution of organic functional groups across the tissue region scanned. Semiquantitation of the concentrations of organic functional groups is performed by comparing values of peak areas obtained at di¡erent regions of the tissue area mapped. In the present study, FT-IR microspectroscopy was used to identify the chemical composition of retinal layers. The layers of the retina have specialized structures that make them distinct. For example, the outer segments contain stacked membrane disks, whereas the inner segments have a high density of mitochondria and ribosomes. The outer and inner plexiform layers have high concentrations of synaptic connections, and the outer and inner nuclear layers contain densely packed nuclei. Thus, the various retinal layers should have distinct chemical compositions related to their specialized functions, and we anticipated that FT-IR microspectroscopy could be used to reveal the chemical composition associated with the various layers. The chemical composition of the retinal layers from albino animals also was investigated.
2. Materials and methods 2.1. Animals Four adult Long-Evans (pigmented) and four adult Sprague-Dawley (albino) rats were kept on a standard light dark cycle with moderate illumination (8 h 210 lux light/16 h dark) which is necessary for normal photoreceptor metabolism. At 5 h into the light cycle, rats were anesthetized, and the eyes were removed, frozen on dry ice, and stored at 370³C. 2.2. FT-IR microspectroscopy In order to obtain a su¤cient signal from low absorbing bands, e.g. 3015 and 1740 cm31 , horizontal frozen sections were prepared at 30 or 40 Wm thick. Two pigmented and two albino rats were analyzed at 30 Wm, and two pigmented and two albino rats were analyzed at 40 Wm. Comparisons between the absorbing bands and various layers were performed on normalized values as described in Table 1 legend. Sections were thaw mounted onto barium £uoride disks, and the layers of the retina were mapped by FT-IR microspectroscopy. Spectra were collected in transmission mode at points on a grid pattern that began in the outer segments and continued across the width of the retina. Mapping experiments at the Microbeam Molecular Spectroscopy Laboratory at Kansas State University utilized an IRWs microspectrometer (Spectra Tech, Shelton, CT, USA). Spectra were collected with a 12 Wm
Table 1 Concentrations of chemical functional groups in di¡erent retinal layers relative to those in the outer nuclear layer 1740 3015 2927 2855 1235 1085 1550 3300
31
cm cm31 cm31 cm31 cm31 cm31 cm31 cm31
OS
OPL
INL
IPL
2.04 þ 0.21 2.63 þ 0.25 1.45 þ 0.10 1.69 þ 0.10 0.62 þ 0.09 0.55 þ 0.09 0.98 þ 0.07 0.74 þ 0.07
1.05 þ 0.10 1.06 þ 0.09 1.14 þ 0.06 1.29 þ 0.09 0.67 þ 0.01 0.66 þ 0.03 0.93 þ 0.05 0.86 þ 0.02
0.96 þ 0.09 0.73 þ 0.09 0.91 þ 0.03 0.97 þ 0.06 0.60 þ 0.02 0.58 þ 0.04 0.88 þ 0.04 0.87 þ 0.07
1.13 þ 0.12 1.08 þ 0.12 1.22 þ 0.08 1.53 þ 0.10 0.39 þ 0.05 0.38 þ 0.05 0.68 þ 0.08 0.55 þ 0.07
The values from two linear maps per functional group were used to obtain the mean value for each function group per region per animal. These values were normalized to the mean values obtained from the outer nuclear layer per each animal. A total of four pigmented animals were used to obtain ¢nal normalized values and standard errors for each functional group. OS, photoreceptor outer segments; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer.
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
U36 Wm projected image plane mask (the long axis was parallel to the length of the retina), 128 coadded scans, and a resolution of 4 cm31 . A background spectrum was collected prior to each mapping experiment with the same conditions, except 256 scans were coadded instead of 128. Additional mapping experiments that were performed at Spectra Tech, Inc., utilized a ContinuWmR microspectrometer equipped with Normarski DIC optics. These experiments utilized a projected double pass (before and after the specimen) single plane mask of 10 WmU30 Wm or 10 WmU60 Wm, 256 coadded scans, and a resolution of 4 cm31 for both specimen and background collection conditions. 2.3. Docosahexaenoic acid Docosahexaenoic acid (DHA) (Sigma, St. Louis, MO, USA) was scanned as a liquid ¢lm on KBr disks at a resolution of 4 cm31 with 256 coadded scans. 2.4. Tissue histology In order to determine whether there was any disparate shrinkage of retinal layers in the frozen sections that were analyzed by FT-IR microspec-
411
troscopy, the histoarchitecture of un¢xed frozen sections was compared to optimally ¢xed and processed Epon sections. Un¢xed, frozen sections, 30^40 Wm thick, adjacent to those analyzed by FT-IR microspectroscopy were thaw mounted onto glass slides and stained by cresyl violet. Epon sections, 1 Wm thick, from retinas ¢xed in 1% glutaraldehyde/1% paraformaldehyde in Millonig's phosphate bu¡er (pH 7.3, 310 mOsm) overnight at 4³C, were stained with toludine blue. 2.5. Statistics Functional groups that were identi¢ed to be concentrated in the outer segments ( s 2-fold above the outer nuclear layer) in pigmented animals (Table 1) were compared between pigmented and albino animals using the unpaired Student's t-test. Signi¢cance was indicated by P 6 0.05. 3. Results 3.1. Preservation of histological structures in un¢xed, frozen retinas Frozen sections are the optimal method of sample
Fig. 1. Stained retina sections. A: A 30 Wm thick, frozen section adjacent to that examined by FT-IR microspectroscopy was mounted onto a glass slide and stained by cresyl violet. The staining reveals that the various retinal layers are preserved in this processing method. B: An optimally ¢xed, 1 Wm thick, Epon-embedded section was stained with toludine blue for comparison with the frozen section. IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer (photoreceptor cell bodies); IS, photoreceptor inner segments; and OS, photoreceptor outer segments.
BBAGEN 24943 1-12-99
412
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
Fig. 2. A series of spectra from the various layers of the pigmented retina. Spectra were collected from photoreceptor outer segments (Spectrum 1), inner segments (Spectrum 2), outer nuclear layer (Spectrum 3), outer plexiform layer (Spectrum 4), inner nuclear layer (Spectrum 5) and inner plexiform layer (Spectrum 6). Note the elevated absorbance values for CNC^H (B), CH2 (C):NH (A), CNO (D) in the outer segments and the elevated absorbance values for PNO (E) and H^C^OH (F) in the outer nuclear layer. The CH2 :NH ratio is also large in the inner plexiform layer.
preparation for FT-IR microspectroscopy since there is no alteration to the native chemical constituents due to ¢xation or tissue processing. The histoarchitecture of stained frozen sections was compared to that for optimally ¢xed Epon sections in order to ascertain the preservation of the retinal layers in the frozen sections. The relative proportions of retina layers were preserved in a similar manner between stained sections from un¢xed, frozen specimens and optimally ¢xed, Epon-embedded sections (Fig. 1).
Thus, the frozen sections that were used for FT-IR microspectroscopy had good preservation of individual retina layers. 3.2. Outer segments of pigmented retina The outer segments had the highest absorbance values of 3015 cm31 (unsaturation based on the CH stretch associated with CNC bonds) relative to all other layers (Figs. 2, 3B). The relative concentra-
C
Fig. 3. Line maps for various chemical functional groups in the retinal layers. The outer segments (OS) contain high absorbance values for 3015 cm31 (CNC^H), 2927 cm31 (CH2 ), 2855 cm31 (CH3 ), and 1740 cm31 (CNO), and low absorbance values for 1235 cm31 (PNO) and 1085 cm31 (H^C^OH). The inner segments (IS) contain high absorptions for 3300 cm31 (N^H) and 1550 cm31 (amide I). The outer nuclear layer (ONL) contains high absorbance values for 1235 cm31 and 1085 cm31 . The ONL and inner nuclear layer (INL) have low absorbance values for 3015 cm31 , 2927 cm31 , 2855 cm31 , and 1740 cm31 . The outer plexiform layer (OPL) and inner plexiform layer (IPL) have absorbance values for 3015 cm31 , 2927 cm31 , 2855 cm31 and 1740 cm31 above those for the ONL and INL but less than in the OS. The two lines represent separate line maps that were collected from tissue areas parallel and adjacent to one another, with no overlap or gap between the areas covered by each line map. A: 3300 cm31 (area points 3658 cm31 to 3134 cm31 , baseline 3734 cm31 to 3116 cm31 ). B: 3015 cm31 (area points 3035 cm31 to 2996 cm31 , baseline 3112 cm31 to 2809 cm31). C: 2927 cm31 (area points 2994 cm31 to 2825 cm31 , baseline 3112 cm31 to 2809 cm31 ). D: 2855 cm31 (area points 2886 cm31 to 2825 cm31 , baseline 3112 cm31 to 2809 cm31 ). E: 1740 cm31 (area points 1774 cm31 to 1730 cm31 , baseline 1800 cm31 ). F: 1550 cm31 (area points 1586 cm31 to 1500 cm31 , baseline 1589 cm31 to 1485 cm31 ). G: 1235 cm31 (area points 1331 cm31 to 1182 cm31 , baseline 1384 cm31 to 1136 cm31 ). H: 1085 cm31 (area points 1135 cm31 to 1011 cm31 , baseline 1135 cm31 to 986 cm31 ).
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
BBAGEN 24943 1-12-99
413
414
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
tion in this layer was V2.6 greater than that in other layers (Table 1). The increased evidence of unsaturation was not simply due to a greater amount of unsaturated lipids in this layer, but also to a higher ratio of double bonds to lipid chain length (Fig. 4B). In addition to unsaturation band intensity, the concentration of carbonyl groups was 2-fold greater in the outer segments compared to the other layers (Fig. 3E, Table 1), and the absorption value at 2927 cm31 also was enhanced, but to a lesser extent. The ratio of CH3 to CH2 was slightly less in the outer segments than in the inner plexiform layers, but greater than in other layers (Fig. 4C). 3.3. Inner segments of pigmented retina The inner segments comprised a narrow layer sandwiched between the outer segments and the outer nuclear layer, and obtaining spectra exclusively from this region was sometimes di¤cult. However, this segment appeared to have pronounced absorptions related to proteins at both 1550 cm31 (Fig. 3F) and 3300 cm31 (Fig. 3A). 3.4. Outer and inner nuclear layers of pigmented retina In the outer nuclear layer the bands at 1235 and 1085 cm31 were highly absorbing (Figs. 2, 3). The CNO and CH2 stretching vibrations were less in-
C
Fig. 4. Line maps for ratios of chemical functional groups. The two lines represent separate line maps that were collected from tissue areas parallel and adjacent to one another with no overlap or gap between the areas covered by each line map. Relative levels of unsaturation via the CNC^H to H^C^OH ratio (A) revealed the greatest evidence of unsaturation in the outer segments followed by the inner plexiform layers, and the lowest values were observed in the outer and inner nuclear layers. The CNC^H to CH2 ratio (B) was greatest in the outer segments. The high peak area for CNC^H in the outer segments indicates a high concentration of double bonds, which is believed to be due to DHA (see Fig. 5). The CH3 to CH2 ratio (C) is lowest in the outer and inner nuclear layers. Outer segments (OS), inner segments (IS), outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL).
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
415
Fig. 5. An infrared spectrum of DHA. The absorption bands for CNC^H (A) and CNO (D) are prominent. Absorption bands for CH3 stretch (C) and CH2 stretch (B) and rock (E) are also evident.
tense in this layer compared to those in the outer segments. The inner nuclear layer had small absorbance values for the CNO and CH2 groups, which were similar in intensity to those observed for the outer nuclear layer. However, unlike the outer nuclear layer, signi¢cant reductions in intensity occurred at 1235 cm31 and 1085 cm31 . The CNC^H:H^C^OH (Fig. 4A) and the CH3 :CH2 (Fig. 4C) ratios were lowest in the outer and inner nuclear layers. 3.5. Outer and inner plexiform layers of pigmented retina The outer and inner plexiform layers showed elevated absorbance values for lipid functional groups. The peak area values for 2927 cm31 , 1740 cm31 , and 3015 cm31 appeared in decreasing order of concentration. This order was opposite to the pattern observed for photoreceptor outer segments. The absolute absorbance for each of these bands was also less than that observed for each band in the photoreceptor outer segments, but was greater than that observed for each band in the outer and inner nuclear layers.
3.6. Albino retina In the albino retina, a selective reduction of organic functional group concentrations was concentrated in the outer segments. In particular, the peak areas of bands at 3015 cm31 , 1740 cm31 and 1235 cm31 were all below 70% of those observed for the normal retina (Table 2). Aside from an increase of the 3015 cm31 band in the inner nuclear layer, no appreciable di¡erences occurred in any of the other functional groups in the various layers between albino and pigmented retinas. 3.7. DHA The absorption bands for CNC^H (3016 cm31 ) and CNO (1713 cm31 ) were very intense in the infrared spectrum of DHA (Fig. 5). The frequency of the carbonyl at 1713 cm31 was attributed to the acid form. Additional bands at 2877 cm31 (CH3 stretch), 2933 cm31 (CH2 stretch), 1432 cm31 (cis CH rock), 960 cm31 (OH wag), and 690 cm31 (cis CH wag) also were prominent. Broadening of the OH stretch of the carboxylic acid group (3400^2500 cm31 ) was evidence of hydrogen bonding.
BBAGEN 24943 1-12-99
416
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
Table 2 Concentrations of chemical functional groups in albino rats relative to pigmented rats OS 1740 3015 2927 2855 1235 1085 1550 3300
31
cm cm31 cm31 cm31 cm31 cm31 cm31 cm31
a
69 69a 81 81 66 84 78 84
OPL
INL
IPL
92 101 100 95 118 118 103 114
86 140 121 112 112 116 110 105
99 102 113 109 110 118 116 122
Data were processed as described in Table 1, but for albino animals. These values were divided by values obtained from pigmented animals and multiplied by 100 to get a percentage. OS, photoreceptor outer segments; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer. a Values obtained from four albino animals were statistically less than those from four pigmented animals.
4. Discussion Fixation and/or processing of tissues can lead to modi¢cations of the chemical constituents of a sample. Since we were interested in analyzing the native chemical composition of retina layers, we utilized un¢xed frozen sections, which provide a better preservation of the innate chemistry than ¢xed tissue. The histoarchitecture of retina was reasonably well preserved by the fresh freezing and cryostat sectioning procedures, since the histoarchitecture observed in frozen sections approximated that for optimally ¢xed Epon sections. Thus, frozen sections permitted sampling of the chemical composition of the individual layers of the retina by FT-IR microspectroscopy. In the rat, the retina is composed predominantly of rod photoreceptor cells. The photoreceptor outer segments contain stacked membranous disks. DHA is a major component of the lipid composition of this layer [4^6] and has been proposed to provide an optimal environment for the conformational change of rhodopsin upon light capture [7]. Because DHA has six CNC bonds [22:6(n33)], it is responsible for the high value of the unsaturation to CH2 ratio and the high absorption value at 3015 cm31 in the outer segments. The infrared spectrum of DHA also revealed an enormous absorbance value for CNO, which accounts for the high concentration for car-
bonyl groups in the outer segments. The relative intensity of bands related to unsaturation and carbonyl groups in the outer segments is considerably higher than that observed for other retina layers and other structures in the brain [2,3,8]. These bands were decreased in albino rats which indicates structural and possible functional alterations in the outer segments in their retinas. A reduction in DHA levels has been observed in light-damaged retinas, and the mechanism of DHA loss is thought to be via light-induced oxidative damage to its CNC bonds [9,10]. Thus, in the albino eye, which is de¢cient in absorbing stray photons, a greater light-induced loss of DHA would be expected compared to that in normal animals. The reduction of chemical functional groups was con¢ned almost exclusively to the outer segments. Thus, the preservation of lipid functional groups in other layers indicates that DHA may be particularly susceptible to light-induced damage because of its six CNC bonds. The reduced amount of DHA in the photoreceptor outer segments of retinas from albino animals compared to those from pigmented animals indicates that various experimental conditions (e.g. [11]) might promote less oxidation in outer segments of albino retinas than in pigmented retinas. Also within the outer segments there was a generalized trend for a slight reduction of other functional groups, which indicates additional changes in this layer. The inner segments contain high densities of mitochondria and ribosomes. The strong absorbance values at 1550 and 3300 cm31 are likely accounted for by these structures, because they contain a relatively high concentration of proteins, but the analysis of this layer without contamination from neighboring structures was a concern. A more detailed analysis of this layer is currently underway with high resolution infrared microscopy using a synchrotron source. The outer nuclear layer contains densely packed cells with a high nuclear to cytoplasmic ratio. The abundance of DNA in this layer is likely responsible for the high absorption values at 1235 cm31 and 1085 cm31 , which are due to PNO and H^O^CH, respectively. Absorptions of H^O^CH and PNO result from the sugar (deoxyriboses) and phosphate (phosphodiester bridges) backbone of DNA. The weaker absorption of these functional groups in the inner nuclear layer probably is due to the lower den-
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
sity of nuclei and a lower nuclear to cytoplasmic ratio. The inner and outer plexiform layers contain a high density of synaptic connections. The membranes associated with synapses, e.g. membrane-bound synaptic vesicles, account for elevated absorbance values for CH2 and CNO functional groups compared to the outer and inner nuclear layers. These values were less, however, than those in the outer segments, which is composed of stacked membranous disks. In summary, the chemical compositions of the various layers were revealed by FT-IR microspectroscopy in retinas from normal and albino rats. The photoreceptor outer segments of retinas from albino rats had a decreased intensity of bands for functional groups found in DHA, which was observed in high concentrations in photoreceptor outer segments of retinas from pigmented animals.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Acknowledgements [9]
This work was supported by NIH NS36544, NS33596, EY 05962 and HD02528. The ¢nal version of the tables and ¢gures was produced by the Image Analysis Facility of the Mental Retardation Research Center at the University of Kansas Medical Center. References
[10]
[11]
417
and mapping of normal brain tissue, Spectroscopy 8 (1993) 40^45. S.M. LeVine, D.L. Wetzel, Analysis of brain tissue by FT-IR microspectroscopy, Appl. Spectrosc. Rev. 28 (1993) 385^ 412. D.L. Wetzel, D.N. Slatkin, S.M. LeVine, FT-IR microspectroscopic detection of metabolically deuterated compounds in the rat cerebellum: A novel approach for the study of brain metabolism, Cell. Mol. Biol. 44 (1998) 15^27. R.E. Anderson, M.B. Maude, W. Zimmerman, Lipids of ocular tissues - X. Lipid composition of subcellular fractions of bovine retina, Vis. Res. 15 (1975) 1087^1090. S.J. Fliesler, R.E. Anderson, Chemistry and metabolism of lipids in the vertebrate retina, Prog. Lipid Res. 22 (1983) 79^ 131. N. Salem, Jr., Omega-3 fatty acids: Molecular and biochemical aspects, I, in: G.A. Spiller, J. Scala (Eds.), New Protective Roles for Selected Nutrients, Alan R. Liss, New York, 1989, pp. 109^228. B.J. Littman, D.C. Mitchell, A role for phospholipid polyunsaturation in modulating membrane protein function, Lipids 31 (1996) S193^S197. S.M. LeVine, D.L. Wetzel, In situ chemical analysis of brain tissue by Fourier transform infrared microspectroscopy, NeuroProtocols 5 (1994) 63^71. R.D. Wiegand, N.M. Giusto, L.M. Rapp, R.E. Anderson, Evidence for rod outer segment lipid peroxidation following constant illumination of the rat retina, Invest. Ophthalmol. Vis. Sci. 24 (1983) 1433^1435. D.T. Organisciak, R.M. Darrow, Y.-L. Jiang, J.C. Blanks, Retinal light damage in rats with altered levels of rod outer segment docosahexaenoate, Invest. Ophthalmol. Vis. Sci. 37 (1996) 2243^2257. M.L. Katz, J.S. Christianson, C.-L. Gao, G.J. Handelman, Iron-induced £uorescence in the retina: Dependence on vitamin A, Invest. Ophthalmol. Vis. Sci. 35 (1994) 3613^ 3624.
[1] D.L. Wetzel, S.M. LeVine, In situ FT-IR microspectroscopy
BBAGEN 24943 1-12-99
Microchemical analysis of retina layers in pigmented and albino rats by Fourier transform infrared microspectroscopy Steven M. LeVine a , Je¡rey D. Radel b , Joseph A. Sweat c , David L. Wetzel a
c;
*
Department of Molecular and Integrative Physiology and the Mental Retardation and Human Development Research Center, University of Kansas Medical Center, Kansas City, KS 66160, USA b Department of Occupational Therapy Education and the Mental Retardation and Human Development Research Center, University of Kansas Medical Center, Kansas City, KS 66160, USA c Microbeam Molecular Spectroscopy Laboratory, Shellenberger Hall, Kansas State University, Manhattan, KS 66506, USA Received 21 June 1999; received in revised form 30 September 1999; accepted 30 September 1999
Abstract Fourier transform infrared (FT-IR) microspectroscopy is a powerful technique that can be used to collect infrared spectra from microscopic regions of tissue sections. The infrared spectra are evaluated to chemically characterize the absorbing molecules. This technique can be applied to normal or diseased tissues. In the latter case, FT-IR microspectroscopy can reveal chemical changes that are associated with discrete regions of lesion sites, which can provide insights into the chemical mechanisms of disease processes. In the present study, FT-IR microspectroscopy was used to analyze sections of retina from normal (pigmented) and albino rats. The outer segments of retinas from pigmented animals were found to have unusually strong absorption values for CNC^H unsaturation and carbonyl functional groups. Docosahexaenoic acid (DHA), a major constituent of lipids in the outer segments, also had particularly high absorption values for these functional groups, which suggests that it is responsible for those enhanced absorption values. Absorbance values for the unsaturation and carbonyl functional groups were substantially reduced in the outer segments of retinas from albino animals. This finding, together with data from other studies on light-induced oxidative events in the retina, indicates a loss of DHA by a light-induced mechanism in albino animals. The outer nuclear layer had strong absorbance values for H^C^OH and PNO functional groups, which is likely due to the sugar phosphate backbone of DNA. The outer and inner plexiform layers were found to contain greater concentrations of CH2 and CNO functional groups than the outer and inner nuclear layers, which is due to the high concentration of synaptic connections in the former layers. In summary, FT-IR microspectroscopy revealed a unique chemical profile in the outer segments compared to other retinal layers, and this profile was altered in albino animals. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Retina; Docosahexaenoic acid; FT-IR microspectroscopy; Outer segment
1. Introduction Infrared spectroscopy is used to identify organic
* Corresponding author. Fax: +1 (785) 532-7010; E-mail: [email protected]
functional groups within a specimen. Fourier transform infrared (FT-IR) microspectroscopy, which combines infrared spectroscopy with microscopy and computer science, allows for chemical analyses to be performed on microscopic regions of samples [1,2]. This technique can be used on tissue sections, and the chemical composition of microscopic regions
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 9 ) 0 0 1 9 5 - 6
BBAGEN 24943 1-12-99
410
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
can be compared to histological structures [1^3]. This chemical information is obtained in the absence of stains or ¢xatives. In addition to probing individual microscopic structures, experiments can be performed that collect spectra along a grid pattern across an area of interest in a tissue section. Interpolation of data between points is used to generate maps that reveal the distribution of organic functional groups across the tissue region scanned. Semiquantitation of the concentrations of organic functional groups is performed by comparing values of peak areas obtained at di¡erent regions of the tissue area mapped. In the present study, FT-IR microspectroscopy was used to identify the chemical composition of retinal layers. The layers of the retina have specialized structures that make them distinct. For example, the outer segments contain stacked membrane disks, whereas the inner segments have a high density of mitochondria and ribosomes. The outer and inner plexiform layers have high concentrations of synaptic connections, and the outer and inner nuclear layers contain densely packed nuclei. Thus, the various retinal layers should have distinct chemical compositions related to their specialized functions, and we anticipated that FT-IR microspectroscopy could be used to reveal the chemical composition associated with the various layers. The chemical composition of the retinal layers from albino animals also was investigated.
2. Materials and methods 2.1. Animals Four adult Long-Evans (pigmented) and four adult Sprague-Dawley (albino) rats were kept on a standard light dark cycle with moderate illumination (8 h 210 lux light/16 h dark) which is necessary for normal photoreceptor metabolism. At 5 h into the light cycle, rats were anesthetized, and the eyes were removed, frozen on dry ice, and stored at 370³C. 2.2. FT-IR microspectroscopy In order to obtain a su¤cient signal from low absorbing bands, e.g. 3015 and 1740 cm31 , horizontal frozen sections were prepared at 30 or 40 Wm thick. Two pigmented and two albino rats were analyzed at 30 Wm, and two pigmented and two albino rats were analyzed at 40 Wm. Comparisons between the absorbing bands and various layers were performed on normalized values as described in Table 1 legend. Sections were thaw mounted onto barium £uoride disks, and the layers of the retina were mapped by FT-IR microspectroscopy. Spectra were collected in transmission mode at points on a grid pattern that began in the outer segments and continued across the width of the retina. Mapping experiments at the Microbeam Molecular Spectroscopy Laboratory at Kansas State University utilized an IRWs microspectrometer (Spectra Tech, Shelton, CT, USA). Spectra were collected with a 12 Wm
Table 1 Concentrations of chemical functional groups in di¡erent retinal layers relative to those in the outer nuclear layer 1740 3015 2927 2855 1235 1085 1550 3300
31
cm cm31 cm31 cm31 cm31 cm31 cm31 cm31
OS
OPL
INL
IPL
2.04 þ 0.21 2.63 þ 0.25 1.45 þ 0.10 1.69 þ 0.10 0.62 þ 0.09 0.55 þ 0.09 0.98 þ 0.07 0.74 þ 0.07
1.05 þ 0.10 1.06 þ 0.09 1.14 þ 0.06 1.29 þ 0.09 0.67 þ 0.01 0.66 þ 0.03 0.93 þ 0.05 0.86 þ 0.02
0.96 þ 0.09 0.73 þ 0.09 0.91 þ 0.03 0.97 þ 0.06 0.60 þ 0.02 0.58 þ 0.04 0.88 þ 0.04 0.87 þ 0.07
1.13 þ 0.12 1.08 þ 0.12 1.22 þ 0.08 1.53 þ 0.10 0.39 þ 0.05 0.38 þ 0.05 0.68 þ 0.08 0.55 þ 0.07
The values from two linear maps per functional group were used to obtain the mean value for each function group per region per animal. These values were normalized to the mean values obtained from the outer nuclear layer per each animal. A total of four pigmented animals were used to obtain ¢nal normalized values and standard errors for each functional group. OS, photoreceptor outer segments; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer.
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
U36 Wm projected image plane mask (the long axis was parallel to the length of the retina), 128 coadded scans, and a resolution of 4 cm31 . A background spectrum was collected prior to each mapping experiment with the same conditions, except 256 scans were coadded instead of 128. Additional mapping experiments that were performed at Spectra Tech, Inc., utilized a ContinuWmR microspectrometer equipped with Normarski DIC optics. These experiments utilized a projected double pass (before and after the specimen) single plane mask of 10 WmU30 Wm or 10 WmU60 Wm, 256 coadded scans, and a resolution of 4 cm31 for both specimen and background collection conditions. 2.3. Docosahexaenoic acid Docosahexaenoic acid (DHA) (Sigma, St. Louis, MO, USA) was scanned as a liquid ¢lm on KBr disks at a resolution of 4 cm31 with 256 coadded scans. 2.4. Tissue histology In order to determine whether there was any disparate shrinkage of retinal layers in the frozen sections that were analyzed by FT-IR microspec-
411
troscopy, the histoarchitecture of un¢xed frozen sections was compared to optimally ¢xed and processed Epon sections. Un¢xed, frozen sections, 30^40 Wm thick, adjacent to those analyzed by FT-IR microspectroscopy were thaw mounted onto glass slides and stained by cresyl violet. Epon sections, 1 Wm thick, from retinas ¢xed in 1% glutaraldehyde/1% paraformaldehyde in Millonig's phosphate bu¡er (pH 7.3, 310 mOsm) overnight at 4³C, were stained with toludine blue. 2.5. Statistics Functional groups that were identi¢ed to be concentrated in the outer segments ( s 2-fold above the outer nuclear layer) in pigmented animals (Table 1) were compared between pigmented and albino animals using the unpaired Student's t-test. Signi¢cance was indicated by P 6 0.05. 3. Results 3.1. Preservation of histological structures in un¢xed, frozen retinas Frozen sections are the optimal method of sample
Fig. 1. Stained retina sections. A: A 30 Wm thick, frozen section adjacent to that examined by FT-IR microspectroscopy was mounted onto a glass slide and stained by cresyl violet. The staining reveals that the various retinal layers are preserved in this processing method. B: An optimally ¢xed, 1 Wm thick, Epon-embedded section was stained with toludine blue for comparison with the frozen section. IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer (photoreceptor cell bodies); IS, photoreceptor inner segments; and OS, photoreceptor outer segments.
BBAGEN 24943 1-12-99
412
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
Fig. 2. A series of spectra from the various layers of the pigmented retina. Spectra were collected from photoreceptor outer segments (Spectrum 1), inner segments (Spectrum 2), outer nuclear layer (Spectrum 3), outer plexiform layer (Spectrum 4), inner nuclear layer (Spectrum 5) and inner plexiform layer (Spectrum 6). Note the elevated absorbance values for CNC^H (B), CH2 (C):NH (A), CNO (D) in the outer segments and the elevated absorbance values for PNO (E) and H^C^OH (F) in the outer nuclear layer. The CH2 :NH ratio is also large in the inner plexiform layer.
preparation for FT-IR microspectroscopy since there is no alteration to the native chemical constituents due to ¢xation or tissue processing. The histoarchitecture of stained frozen sections was compared to that for optimally ¢xed Epon sections in order to ascertain the preservation of the retinal layers in the frozen sections. The relative proportions of retina layers were preserved in a similar manner between stained sections from un¢xed, frozen specimens and optimally ¢xed, Epon-embedded sections (Fig. 1).
Thus, the frozen sections that were used for FT-IR microspectroscopy had good preservation of individual retina layers. 3.2. Outer segments of pigmented retina The outer segments had the highest absorbance values of 3015 cm31 (unsaturation based on the CH stretch associated with CNC bonds) relative to all other layers (Figs. 2, 3B). The relative concentra-
C
Fig. 3. Line maps for various chemical functional groups in the retinal layers. The outer segments (OS) contain high absorbance values for 3015 cm31 (CNC^H), 2927 cm31 (CH2 ), 2855 cm31 (CH3 ), and 1740 cm31 (CNO), and low absorbance values for 1235 cm31 (PNO) and 1085 cm31 (H^C^OH). The inner segments (IS) contain high absorptions for 3300 cm31 (N^H) and 1550 cm31 (amide I). The outer nuclear layer (ONL) contains high absorbance values for 1235 cm31 and 1085 cm31 . The ONL and inner nuclear layer (INL) have low absorbance values for 3015 cm31 , 2927 cm31 , 2855 cm31 , and 1740 cm31 . The outer plexiform layer (OPL) and inner plexiform layer (IPL) have absorbance values for 3015 cm31 , 2927 cm31 , 2855 cm31 and 1740 cm31 above those for the ONL and INL but less than in the OS. The two lines represent separate line maps that were collected from tissue areas parallel and adjacent to one another, with no overlap or gap between the areas covered by each line map. A: 3300 cm31 (area points 3658 cm31 to 3134 cm31 , baseline 3734 cm31 to 3116 cm31 ). B: 3015 cm31 (area points 3035 cm31 to 2996 cm31 , baseline 3112 cm31 to 2809 cm31). C: 2927 cm31 (area points 2994 cm31 to 2825 cm31 , baseline 3112 cm31 to 2809 cm31 ). D: 2855 cm31 (area points 2886 cm31 to 2825 cm31 , baseline 3112 cm31 to 2809 cm31 ). E: 1740 cm31 (area points 1774 cm31 to 1730 cm31 , baseline 1800 cm31 ). F: 1550 cm31 (area points 1586 cm31 to 1500 cm31 , baseline 1589 cm31 to 1485 cm31 ). G: 1235 cm31 (area points 1331 cm31 to 1182 cm31 , baseline 1384 cm31 to 1136 cm31 ). H: 1085 cm31 (area points 1135 cm31 to 1011 cm31 , baseline 1135 cm31 to 986 cm31 ).
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
BBAGEN 24943 1-12-99
413
414
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
tion in this layer was V2.6 greater than that in other layers (Table 1). The increased evidence of unsaturation was not simply due to a greater amount of unsaturated lipids in this layer, but also to a higher ratio of double bonds to lipid chain length (Fig. 4B). In addition to unsaturation band intensity, the concentration of carbonyl groups was 2-fold greater in the outer segments compared to the other layers (Fig. 3E, Table 1), and the absorption value at 2927 cm31 also was enhanced, but to a lesser extent. The ratio of CH3 to CH2 was slightly less in the outer segments than in the inner plexiform layers, but greater than in other layers (Fig. 4C). 3.3. Inner segments of pigmented retina The inner segments comprised a narrow layer sandwiched between the outer segments and the outer nuclear layer, and obtaining spectra exclusively from this region was sometimes di¤cult. However, this segment appeared to have pronounced absorptions related to proteins at both 1550 cm31 (Fig. 3F) and 3300 cm31 (Fig. 3A). 3.4. Outer and inner nuclear layers of pigmented retina In the outer nuclear layer the bands at 1235 and 1085 cm31 were highly absorbing (Figs. 2, 3). The CNO and CH2 stretching vibrations were less in-
C
Fig. 4. Line maps for ratios of chemical functional groups. The two lines represent separate line maps that were collected from tissue areas parallel and adjacent to one another with no overlap or gap between the areas covered by each line map. Relative levels of unsaturation via the CNC^H to H^C^OH ratio (A) revealed the greatest evidence of unsaturation in the outer segments followed by the inner plexiform layers, and the lowest values were observed in the outer and inner nuclear layers. The CNC^H to CH2 ratio (B) was greatest in the outer segments. The high peak area for CNC^H in the outer segments indicates a high concentration of double bonds, which is believed to be due to DHA (see Fig. 5). The CH3 to CH2 ratio (C) is lowest in the outer and inner nuclear layers. Outer segments (OS), inner segments (IS), outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL).
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
415
Fig. 5. An infrared spectrum of DHA. The absorption bands for CNC^H (A) and CNO (D) are prominent. Absorption bands for CH3 stretch (C) and CH2 stretch (B) and rock (E) are also evident.
tense in this layer compared to those in the outer segments. The inner nuclear layer had small absorbance values for the CNO and CH2 groups, which were similar in intensity to those observed for the outer nuclear layer. However, unlike the outer nuclear layer, signi¢cant reductions in intensity occurred at 1235 cm31 and 1085 cm31 . The CNC^H:H^C^OH (Fig. 4A) and the CH3 :CH2 (Fig. 4C) ratios were lowest in the outer and inner nuclear layers. 3.5. Outer and inner plexiform layers of pigmented retina The outer and inner plexiform layers showed elevated absorbance values for lipid functional groups. The peak area values for 2927 cm31 , 1740 cm31 , and 3015 cm31 appeared in decreasing order of concentration. This order was opposite to the pattern observed for photoreceptor outer segments. The absolute absorbance for each of these bands was also less than that observed for each band in the photoreceptor outer segments, but was greater than that observed for each band in the outer and inner nuclear layers.
3.6. Albino retina In the albino retina, a selective reduction of organic functional group concentrations was concentrated in the outer segments. In particular, the peak areas of bands at 3015 cm31 , 1740 cm31 and 1235 cm31 were all below 70% of those observed for the normal retina (Table 2). Aside from an increase of the 3015 cm31 band in the inner nuclear layer, no appreciable di¡erences occurred in any of the other functional groups in the various layers between albino and pigmented retinas. 3.7. DHA The absorption bands for CNC^H (3016 cm31 ) and CNO (1713 cm31 ) were very intense in the infrared spectrum of DHA (Fig. 5). The frequency of the carbonyl at 1713 cm31 was attributed to the acid form. Additional bands at 2877 cm31 (CH3 stretch), 2933 cm31 (CH2 stretch), 1432 cm31 (cis CH rock), 960 cm31 (OH wag), and 690 cm31 (cis CH wag) also were prominent. Broadening of the OH stretch of the carboxylic acid group (3400^2500 cm31 ) was evidence of hydrogen bonding.
BBAGEN 24943 1-12-99
416
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
Table 2 Concentrations of chemical functional groups in albino rats relative to pigmented rats OS 1740 3015 2927 2855 1235 1085 1550 3300
31
cm cm31 cm31 cm31 cm31 cm31 cm31 cm31
a
69 69a 81 81 66 84 78 84
OPL
INL
IPL
92 101 100 95 118 118 103 114
86 140 121 112 112 116 110 105
99 102 113 109 110 118 116 122
Data were processed as described in Table 1, but for albino animals. These values were divided by values obtained from pigmented animals and multiplied by 100 to get a percentage. OS, photoreceptor outer segments; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer. a Values obtained from four albino animals were statistically less than those from four pigmented animals.
4. Discussion Fixation and/or processing of tissues can lead to modi¢cations of the chemical constituents of a sample. Since we were interested in analyzing the native chemical composition of retina layers, we utilized un¢xed frozen sections, which provide a better preservation of the innate chemistry than ¢xed tissue. The histoarchitecture of retina was reasonably well preserved by the fresh freezing and cryostat sectioning procedures, since the histoarchitecture observed in frozen sections approximated that for optimally ¢xed Epon sections. Thus, frozen sections permitted sampling of the chemical composition of the individual layers of the retina by FT-IR microspectroscopy. In the rat, the retina is composed predominantly of rod photoreceptor cells. The photoreceptor outer segments contain stacked membranous disks. DHA is a major component of the lipid composition of this layer [4^6] and has been proposed to provide an optimal environment for the conformational change of rhodopsin upon light capture [7]. Because DHA has six CNC bonds [22:6(n33)], it is responsible for the high value of the unsaturation to CH2 ratio and the high absorption value at 3015 cm31 in the outer segments. The infrared spectrum of DHA also revealed an enormous absorbance value for CNO, which accounts for the high concentration for car-
bonyl groups in the outer segments. The relative intensity of bands related to unsaturation and carbonyl groups in the outer segments is considerably higher than that observed for other retina layers and other structures in the brain [2,3,8]. These bands were decreased in albino rats which indicates structural and possible functional alterations in the outer segments in their retinas. A reduction in DHA levels has been observed in light-damaged retinas, and the mechanism of DHA loss is thought to be via light-induced oxidative damage to its CNC bonds [9,10]. Thus, in the albino eye, which is de¢cient in absorbing stray photons, a greater light-induced loss of DHA would be expected compared to that in normal animals. The reduction of chemical functional groups was con¢ned almost exclusively to the outer segments. Thus, the preservation of lipid functional groups in other layers indicates that DHA may be particularly susceptible to light-induced damage because of its six CNC bonds. The reduced amount of DHA in the photoreceptor outer segments of retinas from albino animals compared to those from pigmented animals indicates that various experimental conditions (e.g. [11]) might promote less oxidation in outer segments of albino retinas than in pigmented retinas. Also within the outer segments there was a generalized trend for a slight reduction of other functional groups, which indicates additional changes in this layer. The inner segments contain high densities of mitochondria and ribosomes. The strong absorbance values at 1550 and 3300 cm31 are likely accounted for by these structures, because they contain a relatively high concentration of proteins, but the analysis of this layer without contamination from neighboring structures was a concern. A more detailed analysis of this layer is currently underway with high resolution infrared microscopy using a synchrotron source. The outer nuclear layer contains densely packed cells with a high nuclear to cytoplasmic ratio. The abundance of DNA in this layer is likely responsible for the high absorption values at 1235 cm31 and 1085 cm31 , which are due to PNO and H^O^CH, respectively. Absorptions of H^O^CH and PNO result from the sugar (deoxyriboses) and phosphate (phosphodiester bridges) backbone of DNA. The weaker absorption of these functional groups in the inner nuclear layer probably is due to the lower den-
BBAGEN 24943 1-12-99
S.M. LeVine et al. / Biochimica et Biophysica Acta 1473 (1999) 409^417
sity of nuclei and a lower nuclear to cytoplasmic ratio. The inner and outer plexiform layers contain a high density of synaptic connections. The membranes associated with synapses, e.g. membrane-bound synaptic vesicles, account for elevated absorbance values for CH2 and CNO functional groups compared to the outer and inner nuclear layers. These values were less, however, than those in the outer segments, which is composed of stacked membranous disks. In summary, the chemical compositions of the various layers were revealed by FT-IR microspectroscopy in retinas from normal and albino rats. The photoreceptor outer segments of retinas from albino rats had a decreased intensity of bands for functional groups found in DHA, which was observed in high concentrations in photoreceptor outer segments of retinas from pigmented animals.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Acknowledgements [9]
This work was supported by NIH NS36544, NS33596, EY 05962 and HD02528. The ¢nal version of the tables and ¢gures was produced by the Image Analysis Facility of the Mental Retardation Research Center at the University of Kansas Medical Center. References
[10]
[11]
417
and mapping of normal brain tissue, Spectroscopy 8 (1993) 40^45. S.M. LeVine, D.L. Wetzel, Analysis of brain tissue by FT-IR microspectroscopy, Appl. Spectrosc. Rev. 28 (1993) 385^ 412. D.L. Wetzel, D.N. Slatkin, S.M. LeVine, FT-IR microspectroscopic detection of metabolically deuterated compounds in the rat cerebellum: A novel approach for the study of brain metabolism, Cell. Mol. Biol. 44 (1998) 15^27. R.E. Anderson, M.B. Maude, W. Zimmerman, Lipids of ocular tissues - X. Lipid composition of subcellular fractions of bovine retina, Vis. Res. 15 (1975) 1087^1090. S.J. Fliesler, R.E. Anderson, Chemistry and metabolism of lipids in the vertebrate retina, Prog. Lipid Res. 22 (1983) 79^ 131. N. Salem, Jr., Omega-3 fatty acids: Molecular and biochemical aspects, I, in: G.A. Spiller, J. Scala (Eds.), New Protective Roles for Selected Nutrients, Alan R. Liss, New York, 1989, pp. 109^228. B.J. Littman, D.C. Mitchell, A role for phospholipid polyunsaturation in modulating membrane protein function, Lipids 31 (1996) S193^S197. S.M. LeVine, D.L. Wetzel, In situ chemical analysis of brain tissue by Fourier transform infrared microspectroscopy, NeuroProtocols 5 (1994) 63^71. R.D. Wiegand, N.M. Giusto, L.M. Rapp, R.E. Anderson, Evidence for rod outer segment lipid peroxidation following constant illumination of the rat retina, Invest. Ophthalmol. Vis. Sci. 24 (1983) 1433^1435. D.T. Organisciak, R.M. Darrow, Y.-L. Jiang, J.C. Blanks, Retinal light damage in rats with altered levels of rod outer segment docosahexaenoate, Invest. Ophthalmol. Vis. Sci. 37 (1996) 2243^2257. M.L. Katz, J.S. Christianson, C.-L. Gao, G.J. Handelman, Iron-induced £uorescence in the retina: Dependence on vitamin A, Invest. Ophthalmol. Vis. Sci. 35 (1994) 3613^ 3624.
[1] D.L. Wetzel, S.M. LeVine, In situ FT-IR microspectroscopy
BBAGEN 24943 1-12-99