raman microscopy and x ray diffraction a combined study of fibrillin-rich microfibillar elasticity

$: Raman microscopy and x ray diffraction a combined study of fibrillin-rich microfibillar elasticity$
Raman Microscopy and X-ray Diffraction, a Combined Study ofFibrillin-rich Microfibrillar Elasticity*

Received for publication, December 17, 2002, and in revised form, June 12, 2003Published, JBC Papers in Press, July 21, 2003, DOI 10.1074/jbc.M212854200

J. Louise Haston‡§, Søren B. Engelsen¶, Manfred Roessle�, John Clarkson**, Ewan W. Blanch‡‡,Clair Baldock§§, Cay M. Kielty§§, and Timothy J. Wess‡

From the ‡Centre for Extracellular Matrix Biology, Department of Biological Sciences, University of Stirling,Stirling FK9 4LA, United Kingdom, ¶Food Technology Centre for Advanced Food Studies, The Royal Veterinary andAgricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark, �European Synchrotron Radiation Facility,B.P. 220, F-38043 Grenoble Cedex, France, the **Department of Chemistry, Joseph Black Building, University of Glasgow,Glasgow G12 8QQ, Scotland, United Kingdom, the ‡‡Department of Biomolecular Sciences UMIST, P. O. Box 88,Manchester M60 1QD, and §§Wellcome Trust Centre for Cell-Matrix Research, Schools of Biological Sciences andMedicine, University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom

Fibrillin-rich microfibrils are essential elastic struc-tures contained within the extracellular matrix of awide variety of connective tissues. Microfibrils are char-acterized as beaded filamentous structures with a vari-able axial periodicity (average 56 nm in the untensionedstate); however, the basis of their elasticity remains un-known. This study used a combination of small anglex-ray scattering and Raman microscopy to investigatefurther the packing of microfibrils within the intacttissue and to determine the role of molecular reorgani-zation in the elasticity of these microfibrils. The appli-cation of relatively small strains produced no overallchange in either molecular or macromolecular microfi-brillar structure. In contrast, the application of largertissue extensions (up to 150%) resulted in a markedlydifferent structure, as observed by both Raman micros-copy and small angle x-ray scattering. These changesoccurred at different levels of architecture and are in-terpreted as ranging from alterations in peptide bondconformation to domain rearrangement. This studydemonstrates the importance of molecular elasticity inthe mechanical properties of fibrillin-rich microfibrilsin the intact tissue.

Microfibrils are essential structural components of the extra-cellular matrix and are widely distributed in both vertebrateand invertebrate tissue, where they impart elastic propertieson all dynamic connective tissues (1–7). Fibrillin-rich microfi-brils have been examined extensively by both x-ray diffractionand electron microscopy techniques. Electron microscopy hasrevealed that microfibrils are macromolecular structures witha regular beaded appearance and a diameter of 10–14 nm. Theaverage periodicity of these beaded microfibrils when isolatedand in the untensioned state is �56 nm, although a range ofperiodicities (33–165 nm) has been observed (4, 6). A funda-mental axial periodicity of 56 nm was also observed by smallangle x-ray scattering techniques, which revealed dominant3rd and 6th orders. This is consistent with the presence of a

regular array of microfibrils with a relative stagger of 56/3 nmand indicates a higher level of order in microfibrillar assembly(8–10).

A number of multidisciplinary studies have been conductedthat investigated the structure-function relationship of fibril-lin-rich microfibrils (4, 6, 8–11). Despite extensive investiga-tion, however, the precise molecular arrangement (and conse-quently the basis of elasticity) of these microfibrils remainsunknown. In the model described by Baldock et al. (4), a seriesof molecular folding events are proposed to occur during theassembly of fibrillin-rich microfibrils. These folds are believedto provide a basis for both the appearance and elasticity ofmicrofibrils. This model is discussed in more detail below. Thisstudy utilized the techniques of small angle x-ray scattering(SAXS)1 combined with parallel Raman microspectroscopy tomonitor changes in both the supramolecular and molecularstructure of fibrillin-rich microfibrils (contained within thezonular filaments of ovine eye) under a variety of differentconditions. The aim of this combined approach was to obtaingreater insight into the basis of fibrillin elasticity.

Small angle x-ray scattering was utilized to characterize anychanges in the supramolecular architecture of fibrillin-rich mi-crofibrils following tissue extension. Any alteration in the pe-riodicity of arrays of laterally aligned microfibrils can be easilymonitored in this way. SAXS is a particularly useful techniqueas it allows the analysis of intact tissue samples in the fullyhydrated state. Previous studies (10) have shown that theapplication of relatively low stresses produced only minor dif-ferences in periodicity. In contrast, further extension of thetissue resulted in a higher periodicity. This study aimed tocharacterize this phenomenon further by monitoring structuralchanges that occur during extension up to the limits of tissueextensibility. Previous work in this area was limited somewhatby the power of x-ray source available and by a low spatialresolution. In this study, the high brilliance beamline ID02 atthe third generation synchrotron European Synchrotron Radi-ation Facility was utilized, which combines a highly collimatedand parallel x-ray beam with a high resolution detector and lowsignal to noise level. The improved resolution and the intensityof X-rays available at this state-of-the-art beamline facilitatedthe investigation of fibrillin-rich microfibrils, both by enablingthe detection of structures with a higher periodicity than pre-

* The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.

§ Supported by Grant 98/S15326 from the Biotechnology and Biolog-ical Sciences Research Council. To whom correspondence should beaddressed. Tel.: 44-1786-467814; Fax: 44-1786-464994; E-mail:[email protected].

1 The abbreviations used are: SAXS, small angle x-ray scattering;cbEGF, calcium-binding epidermal growth factor; TB, transforminggrowth factor-�-binding protein-like domain; EGF, epidermal growthfactor.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 42, Issue of October 17, pp. 41189–41197, 2003© 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 41189

viously seen and by allowing a vastly improved samplethroughput.

In addition to this small angle x-ray scattering work was aninvestigation into possible parallel changes in molecular con-figuration. Raman spectroscopy is widely used in the study ofprotein structure and measures the inelastic scattering of lightdependent upon internal molecular vibrations (12). Particularmolecular groupings give rise to specific frequency shifts, gen-erating characteristic spectra that provide information on pro-tein conformation. Raman spectroscopy can be used to monitorprotein secondary structure, domain-domain configurations,and the local environment of interacting side chains. This tech-nique allowed examination of any changes in molecular confor-mation that arose upon the application of stress to the tissue.

Furthermore, any specific change in molecular configuration(induced for example by temperature changes, covalent modi-fication or intermolecular interaction) can be detected by shiftsin Raman band position. Raman spectroscopy is particularlyuseful for the analysis of aqueous systems as water generatesvery weak Raman scattering, thus producing little interferencewith the macromolecular signal. Raman spectroscopy was thusseen as an ideal method for investigating the role of molecularconformation on the elasticity of arrays of fibrillin-rich micro-fibrils in excised tissue.

The main molecular constituent of fibrillin-rich microfibrilsis the glycoprotein fibrillin, although other molecules are alsoknown to be associated (including microfibril-associated glyco-protein 1 and 2, latent transforming binding proteins and chon-droitin sulfate proteoglycans) (13). It was anticipated that anyalteration in protein conformation upon extension of thesezonular filaments would produce an alteration in Raman spec-tra. These changes may indicate the levels of architecture thatdetermine the elastic response.

EXPERIMENTAL PROCEDURES

Sample Preparation—Zonular filaments were dissected from ovineeye tissue as described previously for bovine tissue (8). Briefly, eyeswere obtained within 24 h post-mortem from a local abattoir. Dissectionof the posterior chamber of the eye produced an intact preparation ofciliary body, lens, and vitreous humor. Zonular filaments within thispreparation were mounted securely on small aluminum frames (�1 �0.5 cm, with a window of �0.7 � 0.3 cm) using cyanoacrylate. Sampleswere kept hydrated using Tris-buffered saline (pH 7.2). Previous workwas conducted using either phosphate-buffered saline (8, 10) or Tris-buffered saline (9). No difference has been found in sample behaviorusing either of these two buffers.

Small Angle X-ray Scattering—Small angle x-ray scattering of ovinezonular filaments was carried out on ID02 at the European SynchrotronRadiation Facility, Grenoble, France, using a 10 m sample to detectordistance. Images were collected over 0.5 s on a Thomson x-ray Intensi-fier (TH 49–427) lens coupled to a FReLoN CCD camera (2048 � 2048pixels). This detector has an active area of size 180 mm and a frame rateof 14 images (1024 � 1024 pixels) per second with a 14 bit nominaldynamic range. The wavelength of X-rays used was 0.09958 nm and thebeam size at the sample was 300 � 300 �m. Three samples wereanalyzed and all exhibited similar properties. Images were obtainedfollowing a variety of different tissue extensions, which were estimatedfrom the overall extension of the macroscopic tissue sample as describedby Wess et al. (10). Briefly, the ends of each tissue sample were fixedand then separated by a distance proportional to the original restlength. Extensions were at �0, 50 (data not shown), 100 (data notshown), and 150%. Images were also obtained following tissue relax-ation back to the original rest length. These particular levels werechosen in order to analyze changes in microfibrillar structure that ariseat the limits of tissue extension, immediately prior to tissue failure.Each sample was preconditioned before analysis to ensure that all hadstable base-line mechanical properties. The level of pre-strain withineach sample was estimated from the extent of arcing of the meridionaland equatorial Bragg reflections. The rest length in this study wasdefined as the length at which the microfibrils could be seen to alignand hence are taking up the macroscopic strain applied to the sample.Tissue extensions were determined subsequent to the measurement of

this rest length. Samples were maintained in a hydrated state usingTris-buffered saline (pH 7.2) and analyzed between thin mica sheets.No beam damage was observed during these very short exposure times.

X-ray Data Analysis—X-ray diffraction two-dimensional imageswere calibrated using rat tail tendon collagen (which indexes on a 1/67nm�1 meridional periodicity) and analyzed using in-house software thathas been used previously to interpret type I collagen diffraction data.An empty cell background has been subtracted from each image. Typ-ical diffraction images can be seen in Fig. 1. Integration of meridionalintensities was carried out using FIT2D (14).

Raman Microscopy—Raman microscopy measurements of ovinezonular filaments were performed on a Labram Infinity dual-laserspectrograph (Jobin-Yvon, Horiba, Lille, France) equipped with a Pel-tier cooled CCD detector. The measurements were carried out using a532-nm frequency doubled Nd:YAG laser (100 milliwatts), a �100 ob-jective, and a 600 g/mm grating resulting in a spectral resolution of �13cm�1. The spectra of a spot size of �4 �m in diameter were acquiredusing an integration time of 60 s. Tissue extensions were performed asdescribed for small angle x-ray scattering. Spectra of tissue in thenative state were collected, following extension to 150% and after re-laxation to the rest length. Spectra are shown in Figs. 2 and 3. Thespectra of intensities were normalized to make direct comparison eas-ier. The quality of the data and the low level of background noise madebackground correction unnecessary.

RESULTS

Small Angle X-ray Scattering—Representative images ob-tained from ovine zonular filaments in the native state, follow-ing tissue extension to �150% of the rest length, and uponrelaxation are shown in Fig. 1. The diffraction patterns ob-tained from the native tissue (see Fig. 1A) correspond to thosereported previously for bovine eye tissue and exhibit a 1st orderaccompanied by a prominent 3rd order in the meridional series(8–10). Following tissue extension to �150%, however, a sig-nificantly different image was produced (see Fig. 1B). In thiscase the meridional series was truncated and the positioning ofthe 1st order moved to a smaller angle. Differences were alsoproduced in the equatorial series, with the extended tissueexhibiting an increased sharpness and indexing on a higherangle. There was also a decrease in arcing of both the meridi-onal and equatorial peaks upon extension. Following relax-ation of the tissue back to the rest length, however, there wasan immediate return to the characteristic, albeit weaker, dif-fraction pattern, with first and 3rd orders visible (see Fig. 1C).This indicates that the extensibility of the tissue was reversibleto this limit. Extension to higher levels resulted in tissue fail-ure. Each of these findings and their implications are discussedin more detail below.

Raman Microscopy—Raman microscopy was performed onovine zonular filaments in the native, extended, and relaxedstates. The spectra obtained were highly reproducible fromsample to sample, and representative spectra are shown in Fig.2. The major peaks have been labeled with the appropriatewave number, and the structural features to which they corre-spond are detailed in Tables I and II (12, 15–18). The spectra ofthe native and relaxed tissue showed a high degree of homologyto each other. This implies that changes in protein conforma-tion, which are induced by tissue extension, are widely revers-ible upon relaxation. This also correlates with the findings fromsmall angle x-ray scattering, which demonstrated a similarsupramolecular arrangement within fibrillin-rich microfibrilsin the native and the relaxed states.

Spectra produced by zonular filaments extended to 150% ofthe rest length, however, were significantly different fromthose derived from either the native or the relaxed tissue (seeFig. 2). Accurate identification of spectral changes that aroseupon tissue extension was achieved by the generation of differ-ence spectra. These were produced by subtracting the intensi-ties of the extended samples from the intensities of the nativesamples. Peaks and troughs on the difference spectra corre-

Fibrillin-rich Microfibrillar Elasticity41190

spond to regions where spectral changes have occurred and alsoallow visualization of the magnitude of these changes. Thesedifference spectra are shown in Fig. 3, and the peaks of interesthave again been labeled.

A number of alterations in Raman spectra occur upon exten-sion of ovine zonular filaments. Changes occur at wave num-bers 815, 831, 855, 921, 939, 1004, 1032, 1085, 1128, 1246,1298, 1319, 1390, 1440, 1580, 1607, 1638, 1654, and 1663 cm�1

(see Figs. 2 and 3 and Tables I and II). Of particular note aredifferences in the amide I and III regions (which reflect proteinsecondary structure) and in the aromatic and aliphatic regions(which provide information on protein main chain structureand domain environments). Each of the changes in Ramanspectra and their relationship to the model of molecular pack-ing within fibrillin-rich microfibrils will now be considered inturn.

DISCUSSION

Small Angle X-ray Scattering—Previous studies (8–10) intothe small angle x-ray scattering of zonular filaments wereconducted using bovine ocular tissue. Legislation within theUK restricting the availability of bovine central nervous tissuehas, however, more recently been extended to also includeocular tissue. Therefore, this study was carried out using ovinetissue as an alternative source of mammalian zonular fila-ments. The x-ray diffraction patterns obtained were remark-ably similar to those observed using zonular filaments ex-tracted from bovine eye. Three samples were analyzed, andrepresentative images are shown here.

The images obtained from the native tissue (Fig. 1A) displayeda series of meridional Bragg peaks, which index on an averagefundamental axial periodicity of 55.4 nm. The patterns exhibiteda strong 3rd order at 18.5 nm, which correlates with previousfindings (8–10) and suggests the staggering of adjacent microfi-brils at a periodicity of one-third of the axial unit cell length.

Following tissue extension of 50%, no significant differencewas observed in the three-dimensional organization of the mi-crofibrils, as observed from the diffraction pattern (data notshown). This also corresponds to previous work that docu-mented only minor changes in the axial periodicity of fibrillin-rich microfibrils upon tissue extension of up to 40 or 50% of therest length (10). In contrast, the application of greater strainsproduced markedly different diffraction patterns. Tissue exten-sion to 150% led to a truncation of the meridional series, withthe 3rd order no longer visible (see Fig. 1B). The loss of adiscernible 3rd order suggests that the existing supramolecularthird staggered array of fibrillin-rich microfibrils has beendisrupted in the 150% extended state. Extension also producedan increase in the fundamental axial periodicity of the micro-fibrils as indicated by a shift in the positioning of the 1st order(corresponding to 103.6 nm). This corresponds to a change of87% in axial periodicity for a tissue extension of 150% andrepresents a non-linear extension of macroscopic and nano-scopic features of fibrillin-rich microfibrils. This finding was

central section of the cylindrical transform). Data were recorded onbeamline ID02 at the European Synchrotron Radiation Facility (cameralength of 10 m). An empty cell background has been removed from eachimage. Strong 1st and 3rd orders indexing on a fundamental axialperiodicity of 55.4 nm are evident in A. B, following extension to 150%,the 3rd order (of the 55.4-nm lattice) is not visible, and the 1st order hasmoved to a lower angle indicating an increased fundamental axialperiodicity. Tissue extension also led to an improved orientation of themicrofibrillar bundles (as indicated by a decreased arcing of the Braggreflections). Following relaxation of the tissue back to the rest length(C), both the 1st and 3rd orders reappear and return to their originalpositions, indicating that the structural changes induced by the appliedstrain are at least in part reversible.

FIG. 1. X-ray diffraction images of ovine fibrillin-rich zonularfilaments before, during, and after tissue extension. The imagesare of the tissue in the native (A); 150% extended (B), and relaxed states(C). The zonular filaments were aligned vertically in the beam. Fromthe beam center the images extend to an R and a Z value of 0.0867 nm�1

(R and Z are, respectively, the radial and axial components of the

Fibrillin-rich Microfibrillar Elasticity 41191

also similar to that reported by Wess et al. (10), where exten-sion to 100% resulted in a shift to an 80-nm periodicity. In thisstudy, tissue extension to around 100% resulted in a periodicityof �82.5 nm (data not shown).

Equatorial reflections in the extended tissue exhibited anincreased sharpness and extended to a higher angle (see Fig.1B), indicating a decreased lateral spacing between individualmicrofibrils under applied strain. Furthermore, both the me-ridional and equatorial peaks became less arced upon exten-sion. This indicates that the microfibrils within the zonulartissue are becoming increasingly aligned along the axis of theapplied force.

Relaxation of the tissue back to the rest length re-establishedweak reflections in the meridional series, similar to those in thenative tissue. A 1st order is visible above the background,indexing on 53.9 nm. A 3rd order is also apparent, occurring at18.2 nm (see Fig. 1C). The recoverable nature of the x-raydiffraction pattern indicates that the characteristic 3rd stag-gered array of fibrillin-rich microfibrils can be re-established.This indicates that any supramolecular disruption that arisesupon tissue extension is reversible and the tissue is elasticwithin the range studied.

These findings correspond with previous reports on the re-versibility of the structural changes that arise following exten-

sion of fibrillin-rich tissues (8–11). In the findings reported byWess et al. (10), it was reported that the characteristic 56-nmperiodicity could be re-established following relaxation of tis-sues extended to 100%. The work presented in this currentstudy demonstrates that following extension to even largerlevels (150%), the original pattern can also be recovered. Thisprovides further support to the role of this tissue reorganiza-tion in microfibrillar elasticity.

Furthermore, studies performed on isolated microfibrils re-ported similar limits of extensibility to those found here (4, 5).Baldock et al. (4) observed two major stable populations ofmicrofibrils within samples extracted from tissue; the first hadperiodicities lower than 70 nm, whereas the second had peri-odicities above 140 nm. Similar results were reported in Erik-sen et al. (5), who also observed two major stable populations ofmicrofibrils within samples extracted from tissue. The firstpopulation had a periodicity of �60 nm, whereas the secondpopulation was found in “highly stretched material” and had aperiodicity of 140–150 nm. It was concluded from these studiesthat microfibrils were reversibly extensible in the range of �56to 100 nm and that irreversible deformation occurred at higherperiodicities (i.e. in the microfibrils of 140–150-nm periodicity)(4, 5).

Both of these studies involved microfibrils that had been

FIG. 2. Raman spectra of ovinezonular filaments before, during, andafter tissue extension in samples A–C.Spectra were obtained using Raman mi-croscopy with a frequency doubled Nd:YAG laser at 532 nm. The spectra of thefibrillin-rich microfibrils in the native tis-sue (i) and in the relaxed tissue (iii) areremarkably similar. In contrast the spec-tra of the zonular filaments following tis-sue extension to 150% (ii) showed a num-ber of differences. This indicates that anydifferences in protein conformation thatarise following tissue extension arelargely recoverable upon relaxation.


isolated from tissue, in contrast to the current study thatexamines the intact tissue. The findings reported here describea reversible elasticity within the intact tissue of up to �100 nm.Isolated microfibrils of these dimensions are reported byBaldock et al. (4) as being reversibly extensible, and as yet,microfibrils of longer periodicity (for example 160 nm) have notbeen detected in the intact tissue. In conclusion, microfibrilsfound within the zonular filaments of mammalian eye tissueare found to be reversibly extensible up to at least 103.6 nm ofperiodicity. This is in accordance with the literature whichreports reversible elasticity in both intact and isolated micro-fibrils at periodicities below 140–160 nm, at which point thedeformation becomes permanent (4, 5, 10, 11).

Raman Microscopy—Raman microscopy provides informa-tion on both long range and local features of protein structure.Consequently, information can be derived about any changesthat occur in both secondary structure and in specific residues.The major alterations that arise upon extension of ovine zonu-lar filaments are in the amide I and III regions and in thearomatic and aliphatic regions (in particular 1410 to 1470cm�1) (see Figs. 2 and 3 and Tables I and II).

Correct interpretation of these data requires consideration ofthe molecular composition of fibrillin-rich microfibrils. Al-though a number of molecules are known to be containedwithin these microfibrils, the dominant component is the mol-ecule fibrillin. Three isoforms of the fibrillin molecule exist,fibrillin-1, fibrillin-2, and fibrillin-3, which have a high degreeof sequence homology. All are large glycoproteins (350 kDa)and possess multidomain structures. The individual moleculeshave an extended, rigid conformation and are �160 nm inlength.

The fibrillin molecule is dominated by 47 epidermal growthfactor-like domains, 43 of which possess calcium-binding po-tential (cbEGF). These are interspersed by 7 transforminggrowth factor-�-binding (TB) protein-like modules, each ofwhich possesses 8 cysteine residues. Fibrillin-1 also contains a58-amino acid proline-rich region near the N-terminal of themolecule. In fibrillin-2 this is replaced by a glycine-rich region;fibrillin-3 has a proline- and glycine-rich region. It has beenproposed that this proline-rich region may serve as a “hingeregion” in molecular folding (4). The molecule also possessestwo-hybrid domains and unique N and C termini, which con-tain furin enzymatic cleavage sites, and are processed pericel-lularly as a prerequisite for assembly. Fibrillin-1–3 also con-tain multiple N-glycosylation sites (6).

EGF-like domains are widely distributed in nature and arefound in more than 70 extracellular matrix proteins. Thesemotifs are composed of between 40 and 50 amino acid residuesand are characterized by a central �-turn and a minor C-terminal �-sheet. 6 cysteine residues interact to produce threedisulfide bonds which stabilize the structure. A subset of thesedomains possess a characteristic amino acid pattern whichconfers calcium binding potential. Calcium has been shown tobind in an N-terminal pocket and is thought to increase therigidity of the domain. A conserved aromatic residue is involvedin stabilizing the binding site (19).

In contrast to the rigid cbEGF domains, the linkage regionbetween the cbEGF domain and the sixth TB module in fibril-lin-1 was found to be relatively flexible (20). Flexibility atTB-cbEGF linkages may therefore contribute to overall molec-ular folding and elasticity (4). The eight-cysteine TB modulesfold into a globular structure composed of �-helices and�-sheets. This structure is also stabilized by four internal di-sulfide bonds.

There are 13 tryptophans and 94 tyrosines in fibrillin-1, allare conserved between bovine and human sequences. One ofthe tryptophans is in the C-terminal domain and would beremoved by furin processing. Of the remaining 12, 1 is in theN-terminal domain, 2 are found in cbEGF domains, and 9 arefound in TB or hybrid domains. In TB and hybrid domains, 8tryptophans are found at a conserved position between the 5thand 6th cysteines. This residue is thought to play an importantstructural role in the center of the hydrophobic core of thisdomain (20). The consensus sequence for the TB domains infibrillin includes a tyrosine or phenylalanine at positions 8 and64 and a tryptophan at position 45.

22 of the 94 tyrosines are found in TB or hybrid domains; 1is in the signal peptide and 6 would be removed upon C-terminal furin processing. Two tyrosines are found in each ofthe N-terminal domains, C-terminal domains, and proline-richregions; five are found in EGF domains, and the remaining 54are found in cbEGF domains. These are predominantly foundat two different conserved positions (42/54). The tyrosine atposition 27 is involved in the calcium-binding site and is part ofthe calcium-binding motif. The tyrosine at position 35 has a keyrole in pairwise domain interactions; the side chain forms hy-drophobic packing interactions with the glycine residue at po-sition 25 in the following domain (see Fig. 4) (19). This inter-action is important in stabilizing domains in their near-lineararrangement.

FIG. 3. Difference spectra of Ramanresults (native and extended for eachsample). A number of peaks and troughsare evident and common to all three dif-ference spectra (A–C), corresponding tostructural differences between the ex-tended and native tissue. These featureshave been labeled with the appropriatewave number. Positive features are indi-cated by numbers at the top of the spec-tra, negative features are indicated bynumbers at the bottom. Tentative assign-ments are given in Table I.


The locations of the amide I (1640–1680 cm�1; carbonylstretching mode) and amide III (1230–1310 cm�1; CN stretch-ing mode) regions of the Raman spectrum in native, extended,and relaxed ovine zonular filaments are detailed in Table II.Significant differences are observed between the tissues, whichcorrespond to different secondary structural features (12, 15).In particular, a major alteration occurs in the amide III regionof the spectrum (Figs. 2 and 3 and Table II). The peak whichoccurs at 1246 cm�1 in the native samples exhibits a decreasedintensity upon extension. This peak is re-established upontissue relaxation. Also of note is the emergence of a relativelyintense peak at 1298 cm�1 within the extended samples. Thispeak is absent in both the native and the relaxed states. Bothof these peaks occur within the amide III region of the spec-

trum and therefore reflect secondary structural features ofproteins. A peak at 1246 cm�1 is usually indicative of thepresence of irregular or disordered chain structure, whereas apeak at 1298 cm�1 usually arises from �-helical structures (seeTable II). The peak at 1298 cm�1 can also contain contributionsfrom main chain CH deformations (12, 21).

Furthermore, these differences are also mirrored in the am-ide I region (1640–1680 cm�1; see Figs. 2 and 3 and Table II).In this region the shift is less obvious; however, the maximumof the amide I peak in both the native and relaxed samplesoccurs at around 1663 cm�1, which is usually concurrent withthe presence of irregular and disordered domains or �-sheet-like structures. Conversely the extended sample exhibits apeak at 1654 cm�1 which generally indicates �-helical content(12). It is important to note, however, that the location of thewater bending mode at �1645 cm�1 can sometimes obscure theamide I region and thus make interpretation more difficult(22).

It is therefore apparent that extension of fibrillin-rich micro-fibrils within zonular filaments induces a conformationalchange, with a decreased occurrence of irregular folds andturns and an increased �-helical content. As discussed in moredetail below, microfibrillar extension may serve to unravelregions of the molecules, leading to an altered overall second-ary structure. These changes are reversible upon tissuerelaxation.

A major change is also observed in Raman spectra between

TABLE IProvisional assignment of Raman peaks in Figs. 2 and 3

Wavenumber Feature Change in intensity(Native-extended)

cm�1

815 1831 Tyrosine, phenylalanine 2855 Non-aromatics (Ile), tyrosine 1880 Non-aromatics (Ile, Val, Thr), tryptophan, tyrosine921 1939 CC deformation, non-aromatics (Lys, Val, Leu) 11004 Phenylalanine 21032 Phenylalanine, tyrosine, tryptophan 11065 CC stretch1085 CC stretch, non-aromatics (Lys, Gln, Ser), phenylalanine 21128 CC stretch, non-aromatics (Ile, Val, Leu, Gln, Ser), tryptophan 21160 Non-aromatics (Ile, Val)1177 Tyrosine, phenylalanine1208 Tyrosine, phenylalanine, tryptophan1246 Amide III (irregular/�-structures) 11271 Amide III, tyrosine1298 Amide III (� helices), CH deformations 21319 CH2 deformations 11343 CH2, CH3 deformations, non-aromatics (Lys, Gln, Ser, Thr), tryptophan, tyrosine1390 Aliphatic, tryptophan, tyrosine 11428 CH2, CH3 deformations, non-aromatic (Gln), tryptophan1440 Aliphatic, tryptophan, tyrosine 21452 CH2, CH3 deformations, non-aromatics (Lys, Ile, Leu)1465 CH2 deformations, non-aromatics (Ala, Ile, Val, Leu), tryptophan, tyrosine1555 Tryptophan1580 Tryptophan 21607 Phenylalanine, tyrosine 21638 11654 Amide I (�-helices) 21663 Amide I (irregular/�-structures) 1

TABLE IILocations of the amide I and III regions

Sample Amide Iwavenumber Feature Amide III

wavenumber Feature

cm�1 cm�1

Native 1663 Irregular/�-sheet 1246 IrregularExtended 1654 �-Helices 1298 �-HelicesRelaxed 1663 Irregular/�-sheet 1246 Irregular

FIG. 4. Schematic representation of a cbEGF pair from fibril-lin-1 (co-ordinates 1EMN) (19). Disulfide bonds are shown in yellow,and the residues involved in the calcium-binding site are colored green.Also shown is the conserved tyrosine residue at position 35 which has arole in pairwise domain interactions with a conserved glycine residue atposition 25 in the following domain (both colored red).


1410 and 1470 cm�1 (see Figs. 2 and 3 and Table I). This regionof the spectra is largely dominated by CH and CH2 deforma-tions, although aromatic residues also contribute. In both thenative and the relaxed tissue the main peak occurs at around1452–1465 cm�1, with a shoulder observed at 1428 cm�1. Inthe extended tissue, however, the peak at 1428 cm�1 disap-pears and the main peak shifts to a lower wave number (be-tween 1440 and 1452 cm�1). A similar change was observed inthe Raman spectra of the polymer isotactic polypropylene inthe conversion from the solid (and partially crystalline) state tothe melt (23). Solid isotactic polypropylene shows a main peakat 1458 cm�1 with a shoulder at 1435 cm�1. Conversely themelt exhibits a single peak at 1435 cm�1. Both peaks were inthat case attributed to aliphatic deformations. The loss of thehigher wave number peak was suggested to correlate with achange in chain conformation that occurs upon melting. Thechange observed in this study is also likely to arise through thealteration in protein conformation induced by tissue extension.The reported change in the amide I and III regions of thespectra indicates a conversion from random coil to �-helices inthe extended state (see above). This conformational change willhave obvious implications for long range chain ordering andinter-chain interactions, as reflected in the altering aliphaticdeformations of these fibrillin-rich microfibrils. These changesappear to be reversible on relaxation of the tissue, as indicatedby the similarity of the relaxed and the native spectra.

Changes in intensity are also observed in a number of otherpeaks, as detailed in Table I. A reversible change is observedupon extension of the zonular filaments to 150% at both 855and 831 cm�1 (see Figs. 2 and 3 and Table I). The presence ofa peak at �850 cm�1 accompanied by a partner at 830 cm�1 is

usually assigned to a tyrosine Fermi doublet. The ratio of theintensities at each of these two wave numbers is commonlyutilized to indicate the degree of hydrogen bonding of the tyro-sine residues. For example, a low value of I850/I830 indicatesthat the –OH group is acting as a strong hydrogen bond accep-tor, whereas a high value generally correlates with donorstatus (12). In a study of the coat protein (pVIII) of the fila-mentous bacterial virus Ff, the absence of a peak at 830 cm�1

was attributed to a highly unusual or hydrophobic tyrosineenvironment (16).

In the results shown here the peak at 831 cm�1 is very weakin both the native and the relaxed spectra. The intensity in-creases slightly upon extension of the samples (see Figs. 2 and3). In contrast, the relatively intense peak at 855 cm�1 exhibitsa decreased intensity upon extension. It is hypothesized thatthis change in relative intensity is therefore caused by largechanges in the hydrogen bonding status of the tyrosine resi-dues within the fibrillin molecules upon extension of the tissue.More than 50% of the tyrosine residues within fibrillin-1 arefound in one of two conserved sites within the abundant calci-um-binding EGF domains. These tyrosines are involved in thecalcium binding consensus sequence and in interactions be-tween adjacent domains. This observed change in tyrosine en-vironment could therefore reflect a change in the environmentof the abundant cbEGF-like motifs.

The peak at 1343 cm�1 contains information on a number ofdifferent structural features, including the environment oftryptophan residues. Again this peak usually forms a doublet,with another peak at �1360 cm�1. The ratio of the two inten-sities (I1360/I1340) provides information on the hydrophilic en-vironment of the tryptophan residues, with a high value indi-

FIG. 5. A model of fibrillin align-ment in microfibrils (adapted fromBaldock et al. (4)). Schematic diagramdepicting a possible folding arrangementof fibrillin molecules in a beaded microfi-bril. A shows the structure of the fibril-lin-1 molecule. Two N- and C-terminallyprocessed molecules associate head-to-tail to give �160-nm periodicity (B). Sub-sequent molecular folding events couldgenerate �100-nm periodicity and then�56-nm periodicity (C and D). Fold sitespredicted to generate periodicity of �100nm are at the N- and C-terminal junctionand the proline-rich region. Fold sites pre-dicted to generate �56-nm periodicity arenot known but may be at two or moreTB-cbEGF junctions within the centralregion of the molecule, including TB3(which has the longest linker sequence).


cating hydrophobicity (12). In this case there is no obvious peakat 1360 cm�1, suggesting hydrophilicity. No change is observedin these peaks upon tissue extension implying that the envi-ronment of the tryptophan residues responsible for this regionof the spectrum is not greatly affected by this tissue extension.Approximately 75% of the tryptophans within the fibrillin-1molecule play an important structural role within conservedsites of the TB domains, close to the �-helical regions (20).These residues are located within the core of the domain, and itis possible that this region of the molecule is relatively resist-ant to any mechanical action. It is important to note, however,that changes are observed in the Raman spectra of a number ofother peaks that contain tryptophan contributions (see TableI). It is therefore possible that the environment of at least someof these residues changes following tissue extension. Further-more, a number of other structural features can also contributeto the peak at 1343 cm�1, including main chain deformations,meaning that definitive changes in tryptophan hydrophilicitymay not be readily apparent from this particular part of thespectra (17).

A greater intensity in the native tissue as compared with theextended tissue is also observed at 939 and 1390 cm�1. Thesepeaks typically provide information on protein side chain envi-ronments (both non-aromatic and aromatic), as well as proteinmain chain conformation (16, 18). The peaks present at 1004,1085, 1128, 1580, and 1607 cm�1 display a greater intensity inthe extended state. These again represent protein side chainand main chain vibrations (12), and the differences in thesepeaks upon extension and relaxation of these fibrillin-rich mi-crofibrils may reflect a changing molecular environment withinthe tissue. These differences also appear to be reversible, due tothe remarkable similarity between the native and relaxed spec-tra. The peaks at 815, 921, and 1638 cm�1 exhibit a fall inintensity upon tissue extension. As yet, however, these peaksremain unassigned.

Therefore, it is apparent that a number of reversible alter-ations occur in the Raman spectra of fibrillin-rich microfibrilsfollowing tissue extension. In particular, mechanical extensioninduces a conformational change with an apparent decrease inrandomly coiled regions of protein and a relative increase in�-helical regions. These changes in secondary structure alsoproduce alterations in the environment of specific residueswithin the microfibrils, as is evident from the observed differ-ences in the aromatic and aliphatic regions of the spectra.These include changes in the hydrogen bonding status of anumber of residues, in addition to differences in main chaininteractions. Upon relaxation it appears that the protein spon-taneously refolds to the original state, indicated by a return tothe native spectrum.

Implications for Fibrillin Elasticity—A number of studies (6,8–11) have demonstrated the reversible elasticity of these mi-crofibrils, although the mechanism of this elasticity is un-known. The precise molecular alignment within fibrillin-richmicrofibrils remains uncertain, although several models havebeen proposed (19, 24–26). The model proposed by Baldock etal. (4) describes a number of folding events which it is hypoth-esized contribute not only to the ‘‘beads-on-a-string’’ appear-ance of these microfibrils but also to their elasticity. This modelis summarized in Fig. 5 and is described briefly below. Forsimplicity, only one row of aligned molecules is discussed; how-ever, in microfibrils there may be 6–8 aligned rows of fibrillinmolecules in cross-section.

In this hypothesis, fibrillin molecules are thought to align ina parallel head to tail fashion. A series of folding events occursthat results ultimately in the generation of a stabilized unten-sioned polymer with a 56-nm periodicity. The inter- and in-

tramolecular stabilizing forces within this structure are be-lieved to account for the elasticity of the microfibril.

In the model, following secretion from the cell, fibrillin mol-ecules are believed to associate through their N and C terminito produce a linear structure with a periodicity relating to themolecular length (�160 nm) (see Fig. 5A). Subsequent foldingis thought to occur within the proline-rich region of the mole-cule, generating a “hinged” structure with a new periodicity of�100 nm (see Fig. 5, B and C). This process allows alignmentof exons 12–15 and 50–64, which aligns known transglutami-nase cross-link sequences (4).

Subsequent folding event(s) are predicted to generate astructure of the characteristic 56-nm periodicity. The linkageregions between the TB and cbEGF domains are relativelyflexible and are thought to contribute to this molecular folding.In particular, the so-called TB3 region (which contains a longlinker region of 19 residues and is located before the centralregion of 12 cbEGF domains) is thought to be especially flexi-ble. From experimental data it is proposed that folding ariseswithin this region of the molecule, as illustrated in Fig. 5D (4).Several other such linkages within the central and C-terminalregions of the molecule may also become folded.

Based on observational studies, it was concluded that indi-vidual isolated microfibrils were reversibly extensible betweenperiodicities of 56 and �100 nm and irreversibly deformed athigher periodicities (up to �160 nm) (4). Due to the correlationof these periodicities with the above model for molecular pack-ing, it was hypothesized that reversible elasticity arisesthrough the unfolding and refolding of the TB3-cbEGF region,producing a new periodicity of between 56 and 100 nm. Furtherextension of the microfibrils is believed to lead to unraveling ofthe proline-hinge region and disruption of the transglutami-nase cross-links. It was therefore proposed that this processcorresponds to an irreversible change in microfibrillarstructure.

This model could be seen to correlate with the observationsmade using small angle x-ray scattering in this study. Minortissue extensions (i.e. up to 50% of the original rest length)produced no significant differences in the axial periodicity ofthe microfibrils. In contrast, application of greater strains (upto 150%) results in the generation of a pattern with a longerperiodicity (103.6 nm). Furthermore, a strong 3rd order is nolonger apparent, implying a disruption to the proposed higherorder staggered array of adjacent microfibrils thought to dom-inate the structure. The characteristic 56-nm periodicity was ineach case recoverable following tissue relaxation. Extension tohigher levels resulted in tissue failure. These observationsdescribe macromolecular changes to physiological bundles offibrillin-rich microfibrils upon application of strain to the sys-tem. It is possible that these effects correlate with those ob-served with isolated microfibrils in vitro (4), with reversibleextension up to �100 nm and the proposed unfolding of theTB-cbEGF domain (see Fig. 5C).

In parallel with the observations made by small angle x-rayscattering, the Raman microscopy of zonular filaments de-scribed here represents the first attempt to correlate the elas-ticity of intact tissues containing fibrillin-rich microfibrils withchanges in molecular conformation and domain-domain inter-actions. Samples examined by both techniques were subjectedto identical experimental conditions to allow direct comparisonof the data. The observations made by Raman microscopy couldalso be seen to correspond to the hypothesis outlined above. Areversible alteration in Raman spectra was observed followingtissue extension to 150%. If the elasticity of the tissue arises atleast in part from molecular unfolding and refolding, thenthese types of changes would be expected as differences arise in


the conformation of the protein.The native tissue had spectra characterized by a relatively

low �-helical content, compared with a higher irregular and�-structural content. In contrast, the extended sample exhib-ited a depleted irregular and �-sheet containing profile and ahigher �-helical content. The abundant calcium-binding EGF-like domains, found throughout the fibrillin molecule, areknown to contain both a �-sheet and a �-turn (19). The TBdomains, which are less prevalent, contain two �-sheets andtwo �-helices (20). The observed changes in this study clearlydemonstrate that a reversible change occurs in the secondarystructure of fibrillin-rich microfibrils following tissue exten-sion. It is hypothesized that this mechanical action leads to anunfolding of the molecules within the microfibrils, as proposedby Baldock et al. (4). The molecular reorganization induced bythis process results in a change in Raman spectra. Microfibrilextension appears to produce an unraveling of certain struc-tural features (in this case the irregular and/or �-turns andfolds), and a relative increase in the proportion of other fea-tures, such as the �-helix, which are normally relatively low inprominence in the fibrillin molecule. Upon relaxation of thetissue, the protein appears to refold into the nativeconfiguration.

A change in molecular configuration upon tissue extension isalso supported by differences in the aromatic and aliphaticregions of the Raman spectra. A large change is observed in thealiphatic CH/CH2 region of the spectrum between 1410 and1470 cm�1. It is proposed that this reflects a change in proteinmain chain interactions induced by the alteration in proteinconformation. A change in tyrosine environment following tis-sue extension is also observed. These residues are prominent inthe cbEGF domains of fibrillin. It is proposed that any molec-ular rearrangement that occurs following microfibrillar exten-sion exposes these domains to different forces and differentparts of the molecule than in the relaxed state. The consequentchange in interaction and environment produces a correspond-ing change in Raman spectrum. In both cases the originalconformation appears to be largely recoverable upon tissuerelaxation, as indicated by a return to the native spectrum.

In conclusion, this study represents the first attempt tointerpret the elasticity of fibrillin-rich microfibrils on both amacromolecular and submolecular level. It is apparent changesoccur in both the packing of the microfibrils and in the confor-mation of the protein constituents of these microfibrils underthe application of large strains (up to 150% tissue extension).The changes that occur appear to be largely reversible uponrelaxation. Further investigation of these structural phenom-ena is warranted to establish in greater detail the structuralchanges that occur and the uniformity of effect. Future work

could include structural investigation of recombinant peptideregions of the fibrillin-1 molecule in an attempt to define moreprecisely the changes observed in this study. Time-resolvedstudies to investigate the rate of change of SAXS and Ramanspectra using a controlled mechanical testing regime could alsobe performed in the future. This would reveal the possibleexistence of intermediate structures and establish the limits ofelasticity, leading to an enhanced understanding of the func-tioning of fibrillin-rich microfibrils.

Acknowledgments—The support of the Centre for Advanced FoodStudies MRI in providing funding and access to the facilities at the Kgl.Veterinær- og Landbohøjskole (Royal Veterinary and Agricultural Uni-versity) is gratefully acknowledged as is the support of the EuropeanSynchrotron Radiation Facility.

REFERENCES

1. Thurmond, F. A., and Trotter, J. A. (1996) J. Exp. Biol. 199, 1817–18282. McConnell, C. J., DeMont, M. E., and Wright, G. M. (1997) J. Physiol. (Lond.)

499, 513–5263. Robb, B. W., Wachi, H., Schaub, T., Mecham, R. P., and Davis, E. C. (1999)

Mol. Biol. Cell 10, 3595–36054. Baldock, C., Koster, A. J., Ziese, U., Rock, M. J., Sherratt, M. J., Kadler, K. E.,

Shuttleworth, C. A., and Kielty, C. M. (2001) J. Cell Biol. 152, 1045–10565. Eriksen, T. A., Wright, D. M., Purslow, P. P., and Duance, V. C. (2001) Proteins

45, 90–956. Kielty, C. M., Baldock, C., Lee, D., Rock, M. J., Ashworth, J. L., and

Shuttleworth, C. A. (2002) Philos. Trans. R. Soc. Lond-Biol. Sci. 357,209–219

7. Kielty, C. M., Wess, T. J., Haston, L., Ashworth, J. L., Sherratt, M. J., andShuttleworth, C. A. (2002) J. Muscle Res. Cell Motil. 23, 581–596

8. Wess, T. J., Purslow, P. P., and Kielty, C. M. (1997) FEBS Lett. 413, 424–4289. Wess T. J., Kielty, C. M., and Purslow, P. P. (1998) J. Cell Biol. 141, 829–837

10. Wess T. J., Purslow, P. P., and Kielty, C. M. (1998) J. Struct. Biol. 122,123–137

11. Wright, D. W., Duance, V. C., Wess, T. J., Kielty, C. M., and Purslow, P. P.(1999) J. Exp. Biol. 202, 3011–3020

12. Miura, T., and Thomas, G. J., Jr. (1995) in Subcellular Chemistry (Biswas,B. B., and Roy, S., eds) Vol. 24, pp. 55–99 Plenum Publishing Corp., NewYork

13. Kielty, C. M., Sherratt, M. J., and Shuttleworth, C. A. (2002) J. Cell Sci. 115,2817–2828

14. Hammersley, A. O. (1993) ESRF Internal Report EXP/AH/93-02, ESRF Infor-mation Office, Grenoble, France

15. Debelle, L., Alix, A. J., Jacob, M. P., Huvenne, J. P., Berjot, M., Sombret, B.,and Legrand, P. (1995) J. Biol. Chem. 270, 26099–26103

16. Overman, S. A., and Thomas, G. J., Jr. (1995) Biochemistry 34, 5440–545117. Overman, S. A., and Thomas, G. J., Jr. (1998) Biochemistry 37, 5654–566518. Overman, S. A., and Thomas, G. J., Jr. (1999) Biochemistry 38, 4018–402719. Downing, A. K., Knott, V., Werner, J. M., Cardy, C. M., Campbell, I. D., and

Handford, P. A. (1996) Cell 85, 597–60520. Yuan, X., Downing, A. K., Knott, V., and Handford, P. A. (1997) EMBO J. 16,

6659–666621. Barron, L. D., Hecht, L., Blanch, E. W., and Bell, A. F. (2000) Prog. Biophys.

Mol. Biol. 73, 1–4922. Lippert, J. L., Tyminski, D., and Desmeules, P. J. (1976) J. Am. Chem. Soc. 98,

7075–708023. Nielsen, A. S., Batchelder, D. N., and Pyrz, R. (2002) Polymer 43, 2671–267624. Reinhardt, D. P., Keene, D. R., Corson, G. M., Poschl, E., Bachinger, H. P.,

Gambee, J. E., and Sakai, L. Y. (1996) J. Mol. Biol. 258, 104–11625. Qian, R. Q., and Glanville, R. W. (1997) Biochemistry 36, 15841–1584726. Trask, T. M., Ritty, T. M., Broekelmann, T., Tisdale, C., and Mecham, R. P.

(1999) Biochem. J. 340, 693–701
