osteogenesis imperfecta: an ray fibre diffraction study

$: Osteogenesis imperfecta: an ray fibre diffraction study$
Annals of the Rheumatic Diseases 1986, 45, 750-756

Osteogenesis imperfecta: an x ray fibre diffractionstudyJ P BRADSHAW AND A MILLER

From the Laboratory of Molecular Biophysics, Department of Zoology, Rex Richards Building, South ParksRoad, Oxford; and the Department of Biochemistry, Edinburgh University Medical School, Hugh RobsonBuilding, George Square, Edinburgh

SUMMARY The use of x ray fibre diffraction to study the molecular architecture of healthy anddiseased human tendon is described. The three dimensional structure of human (finger) tendon isderived to high resolution and is shown to be very similar to that reported for rat tail tendon. Inparticular the presence of the 38 A row line in the diffraction pattern suggests that a high degreeof lateral order within the collagen fibrils is a more widespread feature of tendon tissue than was

formerly realised. Axially projected electron density maps of the 670 A unit repeat of thecollagen fibrils of this tissue, and of tendon tissue from three cases of osteogenesis imperfecta(01), are calculated and compared. The results are in agreement with recent biochemical studiesin suggesting that type I (Sillence) 01 is principally a quantitative, rather than a qualitative,defect of type I collagen biosynthesis. The features by which a molecular lesion may berecognised and characterised from diffraction data are discussed.

Key words: brittle bone syndrome, collagen, synchrotron.

Collagen is the most abundant protein of the humanbody. The collagenous connective tissues are re-sponsible for preserving the structural integrity ofthe body and maintaining the relative positions of itscomponent organs. It is a major constituent oftendon, ligament, blood vessels, fascia, basementand glomerular membrane, and cornea. One unfor-tunate consequence of this widespread distributionof the collagens throughout the body tissues is thatinherited or acquired abnormalities of these proteinsproduce a diverse group of diseases affecting dif-ferent combinations of the connective tissues. To acertain extent the pattern of each disorder reflectsthe tissue distribution of the affected collagen type.A good example of this phenomenon is osteogenesisimperfecta (01), the subject of this study.

Despite being called the brittle bone syndrome 01is not necessarily restricted in its effects to bonetissue but tends to affect all of the type I collagencontaining tissues, to a greater or lesser extent. Thisfact, together with a substantial body of previous

Accepted for publication 12 March 1986.Correspondence to Dr J P Bradshaw, Department of Biochemistry,University of Edinburgh Medical School, Hugh Robson Building,George Square, Edinburgh EH8 9XD.

work, has led to the belief that the widespreaddisruption to the connective tissue systems of thebody caused by OI is primarily caused by anabnormality in the biosynthesis of type I collagen. 1 2Over the past few years it has become clear that

01 is not a single disease but rather a heterogeneouscollection of related conditions. Sillence has classi-fied the syndrome into four main groups accordingto the prevalent features and characteristics ofinheritance of each type.3 4 This paper reports anx ray diffraction study of the molecular basis of thecondition as it occurs in tissue samples obtainedfrom three cases of the milder, dominantly inheritedform of the syndrome. This is the first time that x rayfibre diffraction has been used successfully in thistype of study.

Materials and methods

Tendon samples were obtained from three patientswith type I (Sillence) osteogenesis imperfecta andfrom one control and were kept frozen untilrequired. The frozen samples were dissected toproduce specimens of length 15 mm and diameter0*5-1.0 mm, with the long axis aligned with the axis

750

copyright. on July 12, 2022 by guest. P

rotected byhttp://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.45.9.750 on 1 Septem

ber 1986. Dow

nloaded from

http://ard.bmj.com/

Osteogenesis imperfecta: an x ray fibre diffraction study 751

of the tendon. These specimens were allowed tothaw in physiological Ringer's solution before beingmounted over 0-15 M NaCl in sealed specimen cells.They were stretched to remove the crimp.Two Huxley-Holmes focused beam, x ray cameras

were used, one with a specimen to film length of175 mm, mounted on an Elliot GX13 rotating anodex ray generator, and the other of length 2000 mm,using synchrotron radiation.5 The use of two focallengths allowed over 40 orders of the meridionaldiffraction maxima to be measured from some of thespecimens, and the equatorial and off-equatorialintensity.Each exposure was recorded on a pack of three

films to ensure that the complete range of intensitiescould be measured. The films were processed, andtheir intensity data were digitised by scanning on a

Table 1 Intensities of the meridional x ray diffractionpeaks of tendon collagen from a variety of sources*

Order RTT* Control Oil 012 013

1 1000 1000 1000 1000 10002 27 23 20 11 193 109 98 82 106 914 11 12 11 6 115 54 55 61 63 566 17 17 12 17 217 21 31 37 37 458 6 16 19 19 199 71 66 105 70 7210 19 14 20 20 1711 20 8 12 11 1212 31 29 29 22 2613 2 2 1 3 114 4 7 6 7 715 4 8 10 11 916 5 11 12 13 1317 11 6 5 8 418 4 5 3 2 619 6 3 3 3 320 34 35 33 30 3621 26 28 27 28 3522 11 8 7 13 1023 0 7 3 6 524 0 3 1 3 125 14 19 13 13 1626 5 6 8 5 627 11 8 12 10 928 1 4 3 3 229 9 8 5 7 430 16 13 14 11 1031 1 5 5 7 332 2 4 3 6 433 2 4 5 9. 434 17 12 - - 735 4 5 - - 236 6 5 - - 337 4 9 - - 538 3 6 - - -

*Rat tail tendon; after Hulmes et al.6

Joyce Loebl Scandig 3 microdensitometer with araster setting of 100 Im for the long camera filmsand 50 tim for those exposed in the shorter camera.The digitised data were corrected for backgroundscatter and geometric distortions inherent in theoptical characteristics of the measuring process, thethree films from each pack were scaled to eachother, and the total intensity of each reflection wasintegrated. The integrated intensities of correspond-ing reflections as measured on the two differentcameras were used to scale the two sets of data.

Reciprocal lattice spacings were measured di-rectly off the film with a Nikon profile projectortravelling microscope.

Results

The corrected intensities of the meridional diffrac-tion peaks recorded from diseased and controltendon samples are listed in Table 1. All of the datasets have been scaled to a first order intensity valueof 1000. The justification for this has been givenelsewhere.6

CONTROLSThe x ray diffraction patterns produced fromhuman tendon (Fig. 1) were very similar to thoseobtained from the much studied rat tail tendon7 in anumber of important respects. The low angle,meridional region (corresponding to the axial elec-tron density of the diffracting structure) showed thesame 670 A repeat distance, and the relative

Fig. la



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/

752 Bradshaw, Miller

rig. 1D

Fig. 1 Diffraction patterns ofhealthy human tendon,taken at specimen to film distances of (a) 175 mm and (b)2000 mm. The 38A row line is indicated.

intensities of the first few diffraction peaks werecomparable in the two tissues, indicating that at lowresolution the diffracting structures contain thesame step function of electron density. The equa-torial region of the diffraction patterns (whichcontains the lateral packing information) was alsoextremely similar to the corresponding region of the

rat tail tendon pattern. Some of the control sampleseven produced the 38 A row line, previouslyreported for only a few of the more crystallinetendon types.8 This suggests that the quasi-hexagonal arrangement of collagen molecules pro-posed by Fraser et al7 may be more widespread interms of tendon types than was previously thought.Comparison of the individual meridional intensi-

ties with those recorded for rat tail tendon6 showedthat the two diffraction patterns, and therefore thetwo structures producing the diffraction, were ex-tremely similar to quite a high resolution. Unfortu-nately it was not possible to measure the phaseangles corresponding to these meridional intensitiesbut, since the two types of tendon appear to beremarkably similar in terms of their meridionaldiffraction amplitudes, it is reasonable to assumethat the associated phase angles would not show anymarked differences between the two tissues. Thephases of the first 41 meridional reflections of rat tailtendon have been determined by modelling the axialelectron density of one 'D' repeat of the collagenfibril.6 8 Clearly these phases could also be com-bined with the human tendon intensity data in aFourier synthesis to produce an axial electrondensity model for this type of tendon. The validity ofsuch a procedure is dependent upon the degree ofsimilarity of the two structures. The evidence, asoutlined above, indicates that in most respects thetwo structures are indeed very similar. The calcu-lated Fourier synthesis of the two tissues are shownin Fig. 2a.The scarcity of suitable tissue meant that only one

control sample was used in this study. The justifi-cation for using just one control sample for thiscomparative study may be found in the work ofBrodsky and Eikenberry, who showed in a compara-tive x ray diffraction study of type I collagens from avariety of different sources that whereas collagensfrom different species produce subtle, but repro-ducible, differences in the intensities of the meri-dional diffraction peaks, within any single speciesthe degree of similarity between tendons fromdifferent anatomical locations is very high.9 This isborne out by the fact that type I collagen accountsfor over 90% of the dry weight of tendon tissue, withthe variable proteoglycans constituting only a minorfraction of the total tissue mass.10

OSTEOGENESIS IMPERFECTAAlthough the three samples of OI tissue used in thisstudy came from different patients, and despiteprevious observations that the syndrome is a hetero-geneous group of related conditions,1 2 the diffrac-tion patterns obtained from the three tissue sampleswere extremely similar. Moreover, all three patterns



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/


Fig. 2a

Fig. 2b

were almost indistinguishable from that collectedfrom the control specimen. The Fourier syntheses ofthe meridional diffraction, calculated as for thecontrol samples, also showed strong resemblances tothat of the control (see Figs 2b-d).

Discussion

The use of x ray fibre diffraction methods to in-vestigate connective tissue disease is dependentupon one fundamental assumption: that changes inthe biosynthesis of collagen, of the type brought

about by the diseases studied, will show theirpresence by altering the diffraction patternsobtained from samples of the diseased tissues. Thisposes two important questions. Firstly, in what waysmight such a biosynthetic defect affect the diffrac-tion pattern of collagen? Related to this is thesecond question: how small a change could thetechnique be expected to detect?The well defined, axial, one dimensional crystal-

linity of the collagen fibril, with its repeat distance of670 A, produces an equally well defined reciprocallattice of non-zero points, spaced at intervals of 670



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/


50 100 150 200 Residues

Fig. 2c

50 100 150 200 Residues

Fig. 2d

Fig. 2 Fourier syntheses ofthe axially projected electron density ofcollagen fibrils from a variety ofsources. The ordinaterepresents the electron density ofeach collagen molecule per unit axial translation.

A- 1 reciprocal lattice units along the meridion ofthe diffraction pattern. The width of each reciprocallattice point in any particular direction is inverselyrelated to the number of consecutive unit cells inthat direction which are contributing to the diffrac-tion peak. This number is very large in the axialdirection, and the reflections are therefore welldefined and quite narrow. In the equatorial plane,however, the number is relatively small since

collagen fibrils only measure 1-2000 A in diameter,and it is unlikely that the crystallinity is preservedacross the whole width.11 Each meridional reflectionmay therefore be considered to be disc shaped. Thetendon diffraction pattern is the collective productof a large number of individual fibres, and thelateral spread of the meridional peaks may beexaggerated if the diffracting fibrils do not all sharethe same axis. The effect of this fibrillar disorienta-



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/


tion is to spread the meridional intensities out intoarcs centred on the origin of the reciprocal lattice.This provides the first diagnostic feature of thetendon diffraction pattern, for the size of themeridional peaks and their degree of arcing discloseboth the underlying extent of crystalline order andthe degree of fibrillar disorientation. Thus modifica-tions to the connective tissue components whichdisrupt the molecular architecture of the fibril willbetray their presence in the diffraction patternsproduced by the affected tissues. Disruptions of alarger scale will be shown by an increase in the levelof general background blackening of the film. Theresulting diffraction pattern would have a very poorsignal to noise ratio.

If, while maintaining a high degree of structurein the collagen fibril, a connective tissue diseasemanifests itself by modifying the nature of thatstructure then a corresponding modification will beshown in the diffraction pattern. Changes in thelateral packing of molecules, or in their 'D' stagger,will affect the equatorial or meridional spacingsrespectively, while alterations in the peptide chainconformation or composition will become evidentfrom the intensities of the meridional peaks. Asshown by Hulmes and his coworkers the intensitiesof the first few meridionals are dominated by therelative lengths of the gap and overlap regions of thecollagen fibril.6 Changes in these intensities showaltered proportions of the two regions. Similarly,the non-helical telopeptide extensions are known tohave their prevailing influence on orders in theregion between 8 and 19. The higher orders ofdiffraction contain structural information of a higherresolution. Localised perturbations of the molecularstructure would be expected to affect these intensi-ties in a systematic fashion: an abnormality of thenth order should also be evident in changes to orders2n, 3n, 4n, etc.The second question that needs to be addressed

concerns the resolution of the technique: how smalla structural change could x ray diffraction be ex-pected to detect? In absolute terms the resolution ofthe structural information contained in a diffrac-tion pattern is related to the maximum angle ofdiffracted rays recorded. In terms of axial electrondensity of the collagen fibril the degree of fine detailis limited by the number of meridional intensitiescaptured on the film. In practice, however, thesituation is not so simple for it is necessary to bear inmind two other factors, one a feature of connectivetissue diseases, and one a feature of x ray diffrac-tion. In using x ray diffraction to study connectivetissue disease we are looking for changes at a

molecular level which, when magnified by theinteraction of the defective component with other

similarly affected molecules, are capable of causingthe widespread disruption so characteristic of con-nective tissue disorders. Thus even if the techniqueis not able to detect subtle electron density changesbrought about by a single amino acid substitution itcertainly would be capable of detecting any largerscale disruptions to the molecule or fibril broughtabout by the substitution. Furthermore, it is aninherent feature of x ray diffraction that each part ofthe diffracting structure contributes information toevery part of the diffraction pattern. In other words,even though a structural alteration may be so smallthat its primary contribution to the diffractionpattern occurs outside the maximum angle of datacollection, it might be expected to betray itspresence by its effect upon the rest of the diffractiondata.

In summary, any significant structural alterationto the molecular components of tendon will affectthe x ray diffraction pattern of the tissue in apredictable fashion. Of the three samples of OItissue examined, none showed any systematic modi-fication to the control diffraction patterns. It is there-fore reasonable to assume that the structure of thetype I collagen, which accounts for over 90% oftheir dry weight, is not significantly altered by thedisease. These findings may be explained by one ormore of three options: (a) The type I collagensynthesised by the tendon fibroblasts is perfectlynormal. (b) The collagen produced by the cells isdefective in some minor way, but the defect is toosmall to be detected by x ray diffraction. (c) Thecollagen is a heterogeneous mixture of normal anddefective molecules. The good molecules are as-sembled into fibrils, while the bad ones are discardedat some stage of the biosynthetic process.

Let us consider these options in turn. The first onewould seem to go against the majority of theprevious work on the brittle bone syndrome.1 2 Thesecond option may be subdivided into two possibil-ities. If the basis of type I 01 consists of an alterationin the primary structure of collagen molecules whichbrings about a larger scale disruption of the structureof the fibril or tissue then this explanation would notseem to apply to the tissues examined. As wasexplained above the primary alteration might beinvisible to the techniques employed but the ensuingdisruption would not. This option does also coveranother possibility. The modification to the primarystructure might bring about its effect not bydisrupting the fibril structure but by delaying theproduction of fibrils. The disease would thenmanifest itself by a quantitative rather than aqualitative defect of type I collagen fibrils. Such adefect would not be detectable by present x raydiffraction methods for the fibril structure examined



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/


would appear quite normal. A similar effect wouldensue from the third option, in which the primarydefect is sufficiently substantial to prevent some ofthe collagen precursors from completing theirbiosynthetic pathway.Our data therefore indicate that in the samples

studied the effects of type I 01 are best understoodin terms of a decrease in the total amount of normaltype I collagen fibrils within the connective tissues,rather than a structural defect in those fibrils. Thesefindings are in agreement with previous work on thismilder form of the brittle bone syndrome. Assays ofthe different collagens in skin tissue and fibroblastculture have shown that the ratio of type III collagento type I is increased. 12-14 It has been suggested thatthis might be brought about by a non-functionalallele for the pro-al(I) chain.12 Normal type Icollagen is produced, but at approximately half thenormal rate. The excess a2(I) chains appear tobe degraded intercellularly, there being no evidencefor the formation of hybrids incorporating thesechains, such as al(I).a2(I)2.15 This is essentially thesame scheme as that first proposed by Sykes et al.16

This study serves to illustrate the great potentialoffered by x ray fibre diffraction for the investiga-tion of connective tissue disorders. When comparedwith electron microscopy the method has higherresolution and the sample is maintained in a statewhich closely approximates physiological condi-tions. The data produced contain structural informa-tion at both the resolution of the primary molecularlesion and on the larger scale at which the disruptionis manifest. The forte of x ray diffraction is thedetection of order, and as such it is an ideal tool forthe characterisation of disruptions to the moleculararchitecture of connective tissues by disease.

We thank Dr R J Greenall for assistance with collecting data at

Daresbury SRS and Dr L N Johnson for valuable discussions.

Tissue samples were supplied by Dr M J 0 Francis. This work wassupported by the Arthritis and Rheumatism Council.

References

1 Smith R, Francis M J 0 Houghton G R. The brittle bonesyndrome. London: Butterworth, 1983.

2 Prockop D J, Kivirikko K I. Heritable diseases of collagen.N Engl J Med 1984, 311: 377-86.

3 Silence D 0, Senn A, Danks D M. Genetic heterogeneity inosteogenesis imperfecta. J Med Genet 1979; 16: 101-16.

4 Sillence D 0. Osteogenesis imperfecta: an expanding panoramaof variants. Clin Orthop 1981; 159: 11-25.

5 Nave C, HeUliwell J R, Moore P R, et al. Facilities for solutionscattering and fibre diffraction at the Daresbury S.R.S. J ApplCryst 1985; 18: 396-403.

6 Hulmes D J S, Miller A, White S W, Brodsky Doyle B.Interpretation of the meridional x ray diffraction pattern fromcollagen fibres in terms of the known amino-acid sequence.J Mol Biol 1977; 110: 643-66.

7 Fraser R D B, Macrae T P, Miller A, Suzuki E. Molecularconformation and packing in collagen fibres. J Mol Biol 1983;167: 497-521.

8 Hulmes D J S, Miller A, White S W, Timmins P A, Berthet-Colominas C. Interpretation of the low angle meridi-onal neutron diffraction patterns from collagen fibres in termsof the amino-acid sequence. Int J Biol Macromol 1980; 2:338-46.

9 Brodsky B, Eikenberry E F. Characterisation of fibrous formsof collagen. Methods Enzymol 1982; 52: 127-74.

10 Mathews M B. Connective tissue. Molecular biology, biochemis-try and biophysics. Berlin: Springer, 1975: 19.

11 Miller A, Wray J S. Molecular packing in collagen. Nature1971; 230: 437-9.

12 Barsh G S, David K E, Byers P H. Type I osteogenesisimperfecta: a non-functional allele for proalpha 1(I) chains oftype I procollagen. Proc Natl Acad Sci USA 1982; 79: 3838-42.

13 Pope F M, Nicholls A C. Mainz symposium on bone dysplasias.1982. (Abstract).

14 Rowe D W, Poirier M, Shapiro J R. Osteogenesis imperfecta: agenetic probe to study type I collagen biosynthesis. In: VeisA, ed. Chemistry and biochemistry of mineralised connectivetissues. Amsterdam: Elsevier, 1981: 155-62.

15 Barsh G S, Byers P H. Reduced secretion of structurallyabnormal type I procollagen in a form of osteogenesis imper-fecta. Proc Natl Acad Sci USA 1981; 78: 5142-6.

16 Sykes B C, Francis M J 0, Smith R. Altered relation of twocollagen types in osteogenesis imperfecta. N Engl J Med 1977;296: 1200-3.



j.com/

Ann R

heum D


ber 1986. Dow

nloaded from

http://ard.bmj.com/

osteogenesis imperfecta: an ray fibre diffraction study

Documents