Wilson

Low-angle X-ray diffraction studies of reduced and silver-stained -keratin G. A. Wilson* Textile Physics Laboratory, Department of Textile Industries, The University, Leeds 2, UK (Received 31 March 1971 ; revised 11 June 1971)

When the cystine in Lincoln wool fibres is reduced and the fibres subsequently 'stained' with silver nitrate, the low-angle X-ray diffraction pattern becomes considerably enhanced, and to a good approximation represents the regular distribution of silver atoms throughout the fibre. This regular distribution of silver atoms is equivalent to the distribution of cystine, since if the reduced cystine is Jolocked by alkylation before staining, the silver uptake is low and the only enhancement which occurs is a very small intensification of the low-angle equatorial reflections. The precise nature of the enhanced diffraction patterns varies considerably with the prior reduction treatment and it is possible to distinguish between matrix and microfibril fractions of cystine. When the microfibril cystine is labelled the low-angle meridional reflections contract in spacing and this is accompanied by a similar change in fibre length. There is no apparent change in the high-angle, helical pattern, however, which leads to the conclusion that the ordered keratin protein chains are made up of alternating helical and non-helical sections. A model for the matrix is also proposed, consisting of short, randomly orientated c~-helical segments, which become orientated about the fibre axis where they form the surfaces of the microfibrils. A possible mode of interaction between the matrix and microfibrils is also suggested.

INTRODUCTION The low-angle X-ray diffraction of ~-keratin impregnated with heavy metals after reduction of the disulphide bonds has been reported by several workers 1 4. Both the equatorial reflections (at approximate Bragg spacings of 80/k (el), 45/k (e._,) and 27• (ea))and the meridional reflections (as orders of a macroperiod of 198,~) have been shown to be enhanced. The equatorial patterns of wool fibres obtained using silver nitrate and phenyl mercuric hydroxide as the heavy metal reagents have been related to the histological distribution of the stains throughout the fibre 3. The matrix-microfibril contrast observed in the transmission electron microscope 3, 6 has suggested a line structure of 'holes' arranged in an electron-dense medium '~ and this structure has been related to the spacings and intensities of el, e2 and e3. The 'holes' have been equated with the * Present address: School of Biological Sciences, The University, Bradford 7, Yorkshire, UK.

microfibrils, and hence it has been concluded that the stain is entering, preferentially, the matrix. The fact that the observed effects take place after prior reduction of the disulphide bonds to thiol groups suggests that the binding site for the heavy atom is at a sulphur atom, and thus a cystine-rich matrix is implied. The interpretation of the changes in the meridional X-ray diffraction patterns produced by heavy atoms has been made difficult by the fact that previous workers have not only used different metals but also different redmction treatments prior to staining. This paper examines the effect of difl'erent prior reduction treatments on the X-ray diffraction of silver-stained keratin. EXPERIMENTAL Lincoln wool samples were prepared in which (i) 40~o of the wool cystine was reduced with thioglycollic acid (termed 'most reactive' cystine); (it) the 'most reactive'

POLYMER, 1972, Vol 13, February

63

X-ray diffraction studies o f ~-keratin : G. A. Wilson

cystine was reduced as above, alkylated with ethylene dibromide, and the remaining 'least reactive' cystine reduced with tetra(hydroxy methyl) phosphonium chloride (THPC); (iii) 80~o of the cystine was reduced with T H P C ; and (iv) all the cystine was reduced and alkylated with ethylene dibromide by two reductionalkylation cycles (using thioglycollic acid and THPC as the two respective reducing agents). The samples were all stained in solutions of silver nitrate, and X-ray photographs were taken of the washed and dried samples. Parallel experiments were carried out on single fibres and tests were made for fibre contraction at each stage of the reaction sequence. Lincoln wool was prepared for analysis by Soxhlet extraction followed by a dilute (6 × 10-4 N) hydrochloric acid wash to remove any metal ions and reduce the ash content of the wool 7, and finally washing in conductivity water to the isoelectric point of wool (pH = 5.1). Low-angle X-ray photographs were taken using a 'pin-hole' camera designed in this laboratory s. Nickel filtered CuK~ (,~= 1.54A) radiation was used and was generated from a Hilger Y-25 100t~m microfocus tube operating at about 35 kV and 2.5mA. High-angle photographs using CuK~ radiation were taken on simple glass-capillary collimated cameras, also designed in this laboratory 9. The source of X-rays was a Hilger and Watts Y-90 X-ray generator, working at 34kV and 14mA. A nickel filter was employed as before, and silver foil was inserted in the centre of each sample in order to measure the exact reflection spacings. The X-ray photographs were analysed on a Joyce Loebl Automatic Recording Microdensitometer Model MkIIIC. All traces were corrected for background scatter by comparison with a trace at an angle of 45 ° to the equator. The reduction procedure with 0"45M thiogtycollic acid followed by a standard wash to yield samples of known constant thiol content (equivalent to 4 0 ~ reduction), and the silver nitrate staining procedure at pH 5-4, were both developed in this laboratory 3. Uptakes of silver were measured gravimetrically against blank samples. Reduction with 4 ~ (w/v) THPC at pH 5.410 yielded a sample with 80~o of the cystine converted to thiols. In the case where the sample had already been reduced by thioglycollic acid and subsequently alkylated, all t h e remaining cystine was reduced. Before further reaction THPC reduced samples were quickly washed in successive changes of conductivity water for 10 rain. The alkylation procedure with 0.05 M ethylene dibromide at pH 8 has been described elsewhere al. The thiol contents of all the samples were measured by gravimetric uptake of phenyl mercuric hydroxide3, and by measurement of the remaining cystine after alkylation of the thiols, by the Shin O'Hara method 12. This was compared with the cystine content of untreated wool, which was found to be 860 t~mol/g, in agreement with the analyses of other workers ~, 13.

Table 1 Uptakes of silver after reduction of cystine in Lincoln wool

Reduction treatment 0.45M thioglycollic acid (pH 6) 0.2M THPC (pH 5.4) 0.45M thioglycollic acid (pH 6) alkylation with (CH2)2Br2 further reduction with 0.2M THPC (pH 5.4) Complete reduction and alkylation

Thiol content (Fmol/g wool)

Silver uptake (t~mol/g wool)

340 700

1220 2120

580

1470

0

310

Table 2 Intensity distribution of the low-angle equatorial reflections el, e2, e3 at approximate respective Bragg spacings 80 ~., 42 ~, 27 ~ in silver-stained Lincoln wool

Treatment

lel : le~ : le3

Approximate enhancement (intensity ratio le2 : I~.s~)*

Untreated sample 'Most reactive' cystine silver stained 'Least reactive' cystine silver stained 80% of cystine silver stained All of cystine reduced and alkylated before silver staining

10 : 0.75 : 1

0-4 : 1

14 : 3.75 : 1

50 : 1

3 : 1.9 : 1 14.5 : 3 : 1

50 : 1 50 : 1

14 : 1.2 : 1

1 :1

Intensity ratio

* Intensity of e2 compared with that of the main equatorial highangle helical reflection at a Bragg spacing of 9.8 ~,

far too high to be accounted for by a stoichiometric 1 : l reaction with thiols. Most of the reaction appears to take place with these groups, however, as the introduction of an alkylation step between the reduction and staining reactions considerably decreases the metal uptake. The low-angle X-ray diffraction patterns of the samples, with the exception of that of the reduced and alkylated sample, become considerably intensified with respect to the high-angle coiled coil diffraction after staining (Table 2). To a good approximation, therefore, they describe the distributions of silver atoms in the different samples. The low-angle diffraction pattern of the reduced and alkylated sample, however, shows no enhancement along the meridian after staining (densitometer traces showed the meridional pattern to be identical to that obtained from untreated Lincoln wool), and very little along the equator (Table 2). Thus the enhanced diffraction patterns may be assumed to be due to silver atoms located at cystine residues, and hence describe the distribution of cystine throughout Lincoln wool fibres.

Low-angle X-ray diffraction of the reduced and stained samples RESULTS AND DISCUSSION

Specificity of silver nitrate for reduced cystine The silver uptakes for all the samples considered are shown in Table 1. It should be noted that the uptakes are

64

POLYMER,

1972, Vol

13, F e b r u a r y ,

The low-angle X-ray diffraction pho.tograph of the sample in which the 'most reactive' cystine has been reduced and labelled with silver is shown in Figure 1, and that of the sample in which the 'least reactive' cystine has been reduced and labelled is shown in Figure 2.

X-ray diffraction studies of =-keratin: G. A, Wilson

8th

ii7

~ii!!i~

Figure 1 Low-angle X-ray diffraction photograph of Lincoln wool that has been silver stained after reduction of the 'most reactive' cystine

- ~th

lower the intensity of the external interference giving rise to el. By drawing an analogy with untreated Lincoln wool, where le2 : l e a ~ l : 1, the observed lowered intensity ratio from 4 : 1 to 2 1 may be similarly explained. The theory explaining the low-angle equatorial diffraction of untreated Lincoln wool is based on the scatter from regularly distributed protofibrils (3-strand coiled coils) inside the microfibri114. In the usual stained specimens as typified by Figure 1, the electron density difference between the matrix and microfibrils is so large that the arrangement of protofibrils inside the microfibrils becomes unimportant and may be neglected. However, if the stain enters the microfibrils in large quantities, the density difference between the microfibrils and matrix will become equalized, and essentially the same diffraction conditions as those for untreated Lincoln wool will occur, i.e. the low-angle equatorial diffraction will be based largely on the arrangement of protofibrils. The effect of this in the stained sample will presumably be to lower the intensity ratio lee : le3. these results suggest, tneretore, that the 'most reactive' cystine is located mainly in the matrix, and the 'least reactive" cystine in the microfibril.

th rd

Figure 2 Low-angle X-ray diffraction photograph of Lincoln wool that has been silver stained after the 'most reactive' cystine has been blocked and the 'least reactive' cystine reduced

Important differences are apparent, both along the equator and the meridian. These are discussed in detail below. The equatorial diffraction. The intensities of the lowangle equatorial reflections are summarized in Table 2. The equatorial diffraction of the sample in which the 'most reactive' cystine has been reduced prior to staining is identical to that found by other workers ~, a This type of pattelm, taken in conjunction with the observation in such samples of matrix-microfibril contrast in the transmission electron microscope, is generally interpreted as being produced by scatter fi'om a system of 'holes' (microfibrils) embedded in an electron dense (i.e. silver dense) medium (the matrix). This type of scatter is characterized by a high intensity of el due to external interference between microfibrils, and e,, and ea representing the scatter from a single cylinder, radius 37'5A, having an intensity ratio le,, : l e a ~ 4 : 1. When, however, the 'least reactive' cystine is labelled, the relative intensity of et becomes much lower, and le., : lea ~ 2 : 1. These differences may be explained, qualitatively at least, by postulating that when the 'least reactive" cystine is reduced and stained, considerable quantities of silver enter the microfibril. The effect of this will be to randomize the histological distribution of silver, and hence

The meridional diffraction. When the 'most reactive' cystine is labelled the 8th order meridional reflection alone appears to be notably enhanced (Figure 1). When, however, the 'least reactive' cystine is labelled the 3rd, 5th a n d 8th orders are enhanced, with the 3rd order particularly strong (Figure 2). This last feature of the meridional difl'raction confirms the results of workers using mercury compounds as the stains ~,4 and provides a simple explanation for the non-correlation of their results with those of workers using silver compounds. Evidently the prior reduction treatment is extremely important in determining which reflections are enhanced on staining. The 5th order enhancement has also been reported previously, but this has been associated with a lysine repeaO ~', HL It is quite possible that this repeat is present in addition to that attributed to cystine in the present paper. The intensity ratios of the meridional reflections of all three enhanced diffraction patterns are shown in Table 3. It will be noted that when enhanced, the intensity ratio of the 3rd and 5th orders is constant, but the relative intensity of the 8th order varies with the exact chemical treatment, being strongest when the 'most reactive' cystine has been labelled, and weakest when the 'least reactive' cystine has been labelled. It would appear that the intensity of the 8th order is associated with 'most reactive' (or matrix) cystine, and the 3rd and 5th orders with 'least reactive' (or microfibril) cystine.

Table 3 Relative intensities of enhanced meridional reflections in silver-stained Lincoln wool 1

Treatment

Intensity ratio lard : /5th : /Sth

'Least reactive' cystine silver stained 80% cystine silver stained 'Most reactive' cystine silver stained

10:2:1.5 10:2:4 10" : - - : >10

* I n this sample the 5th order is very weak and has not been measured. Also there appears to be very little enhancement of the 3rd order

POLYMER,

1972, Vol

13, F e b r u a r y

65

X-ray diffraction studies of =-keratin: G. A. Wilson Table 4 Bragg spacings of the meridional reflections enhanced when the 'least reactive' cystine in Lincoln wool is silver stained Bragg spacing (~) Order

Untreated Lincoln wool

'Least reactive' cystine silver stained

3rd 5th 8th Macroperiod (,&,)

65.8 39* 24.4 198

61.65 36.80 23.20 185

* A very weak reflection which is only occasionally observed in untreated specimens s,~-~

The shapes of the 3rd and 5th order reflections are different (Figure 2). The 3rd order, like the 8th, is a sharp meridional arc, but the 5th order splits into three distinct reflections which form a rudimentary layer line. In the light of work on collagen iv, and more recently on untreated a-keratin 8, this would suggest that the microfibril cystine itself is subdivided into cystine in two different structural positions (although of similar reactivity to reducing agents). When only the 8th order reflecUon is enhanced, its Bragg spacing is the same as in untreated specimens at 24.4A. When the 3rd, 5th and 8th orders are enhanced, however, the spacings all contract by 6.5~ and the macroperiod falls from 198 to 185/k (Table 4). This change in low-angle spacing is accompanied by a contraction in fibre length of 6.3 % (average of five single fibres) which occurs only at the final silver staining stage. No contraction is observed for samples which do not show spacing changes. Thus there is a direct correlation between change in X-ray spacing and change in fibre length. As the contraction in length occurs at the final silver staining reaction, it would suggest that the contraction takes place in the parts of the protein chain where the silver is being attached, rather than continuously along its length. This contention is supported by a study of the high-angle pattern. The high-angle X-ray diffraction of keratin is generally accepted as being that given by a 2- or 3-strand coiled coiP 8, xg. (Recently, other a-helical forms have been shown to be possibilities 2°-23, but, this is irrelevant to the following argument.) In particular, the strong meridional ~eflection at 5.15A represents the pitch of an a-helix which has been twisted to form one strand of a coiled coil. A regular contraction would presumably cause a decrease in the spacing of the reflection (for example, compare the fact that when wool is extended by 20 ~, the 5.15A reflection increases in spacing by 2 ~ 24) but accurate measurement showed that the reaction sequence had caused no change. However, further considerahon of the high-angle pattern, and this reflection in particular, suggests that the implied irregular coiled coil contraction is not taking place either. There is no sign of any disorder which would arise from an irregular contraction of the coiled coil. Although to establish the fact completely it would be necessary to measure quantitatively intensity changes in the 5"15/~ reflection which result from silver staining-a very difficult task owing to the presence of the silver swamping the natural diffraction pattern--the results do suggest that there is no contraction in the coiled coil, 66

P O L Y M E R , 1972, Vol 13, February

helical regions of the fibre. This, in turn, leads to the deduction that the contraction is taking place in nonhelical regions which alternate in series with the coiled coil (or other s-helical regions). If this is so, it may reasonably be assumed that the non-helical regions are stabilized to a large extent by disulphide bonds, and the cleavage of these bonds by reduction, and the subsequent reaction with silver nitrate, cause these regions to contract. The very great intensity of the 3rd order reflection suggests that the major repeat of one helical plus one non-helical region is 66A. This agrees with analyses of other workers studying the 3rd order reflection in untreated keratin s, 25, who suggest that it arises from a major discontinuity at this spacing. This discontinuity has been linked by these workers with the anti-helical amino acids 25 or, more specifically, cystine residues 8. The present results are in full accord with the latter postulate. The much weaker 5th order reflection (suggesting a 39.5A repeat) is not considered to be associated with helical-non-helical discontinuities. However, as it always has the same comparative intensity as the 3rd order, it appears to .represent a repeat which is part of the same chain. It is difficult to envisage how regular contractions at 66A intervals could bring about a regular contraction of the 39.5A spacing. One possibility is that in untreated specimens this repeat is slightly irregular. The reflection is extremely weak and does not index satisfactorily on a macroperiod of 198A, its actual Bragg spacing being recorded as 39A s,2~. These characteristics would be expected if the cystine residues giving rise to the repeat were occasionally spaced at intervals shorter than the theoretical 39-5 A. The contraction of non-helical regions every 66A could then be envisaged as making the spacing more regular, as well as decreasing its overall value. This explanation would also account for the fact that other workers who have recorded intensification of the 3rd order after labelling cystine with mercury compounds but no change in spacing have not reported any intensification of the 5th order 2,4.

CONCLUSIONS

The structure of the matrix The broad halo (Bragg spacing=25.~) observed in X-ray diffraction studies of silver-stained fibres and the sharp diffraction ring (Bragg spacing= 21 ,~) observed in X-ray diffraction patterns of the silver-stained keratogenbus zone of developing fibres have been interpreted as indicating matrix order2V, 28. In the latter studies 2s electron micrographs revealed the presence of silver globules of the same order of size, between the microfibrils, and it was also reported that after keratinization but before dehydration the diffraction ring became broader. These results were interpreted in terms of the matrix originally being a globular entity which becomes distorted on keratinization owing to oxidation of cysteine to cystine. More recently, very wide angle studies have suggested that the matrix consists of short, randomly orientated s-helical segments 3, s, and the 25,~ halo has been related to the leflgth of a segment 8. A closer examination of this

X-ray diffraction studies of a-keratin : G, A. Wilson 200

d:D

I C 0 C

The structure of the protofibrils

, 150

o

l/

I00o<

50 X-X X -

9 x ~/ tb- ~

associated in the present work with a regular matrix repeat, could arise from complete orientation of the matrix segments about the fibre axis. This would most likely occur at the surfaces of the microfibrils (a possibility suggested by early workers on the X-ray diffraction of silver-stained keratin1), to enable maximum disulphide crosslinking between the matrix and microfibrils.

b

O

Correlation of the 'least reactive' cystine with microfibril cystine and co-ordination of the data obtained from the diffraction pattern of the sample in which this fraction has been labelled with silver lead to a general model for the microfibril sub-units or protofibrils (Figure 3b). The helical regions in this Figure are shown as 3-strand coiled coils, and the non-helical regions are shown to have a finite length, so that they contain some cystine residues which form part of the 39A repeat. This has been done because it is well known that cystine is not readily incorporated into a-helical structures. it is possible that the non-helical regions tbrm a fairly large proportion of the protofibrils. Optical transform results show that 607k lengths of coiled coil will yield well-ordered helical reflections, although if the length is reduced to 27A they become diffuse e'5. The idea of alternating helical and non-helical regions has been the basis of various mechanical models e9, a0, and also some recent models derived from chemical studies al-aa. It represents a significant departure, however, from recent X-ray diffraction models such as the segmented rope model, where the coiled coil is continuous and the coiling takes place at specific residues2a, ~5, and the straight ~-helix models ~°-'~. These models have been derived because the experimental data from diffraction patterns do not fit exactly the theoretical continuous coiled coil diffraction. Models similar to that presented in this paper may explain the discrepancies equally well.

~

The unified structure Figure 3 The molecular configuration of c~-keratin. (a) Proposed structure for the matrix, consisting of short units of helical plus non-helical material, about 25 ~, long. Cystine (half-cystine represented by © ) i s located in the non-helical segments and the disulphide bonds between the units produce a 'rigid' structure. The units are orientated randomly except where they form the surface of the microfibril. Here they become orientated along the fibre axis. (b) Shows one protofibril on the periphery of a microfibril. It consists of a 3-strand coiled coil, interspersed at regular intervals by non-helical sections. The basic unit of one helical plus one non-helical segment is 66,~ long. The non-helical sections all contain cystine and there is a 66 ~ cystine (halfcystine represented by ©) repeat. The other cystine repeat (half-cystinerepresented by x ) is shown as 39.5~ but may be irregular. This cystine may be located either in the coiled coil or non-helical segments. Possible disulphide links with the matrix to form 'rigid' parts of the structure are shown

halo has revealed it to be more intense and ol slightly greater spacing along the meridian s , suggesting the segments to be partly orientated about, and extended along the fibre axis. These deductions are in full agreement with a recently derived mechanical model for the matrix 29. In th~s type of model (Figure 3a) it is easy to envisage how the 25A meridional reflection, which has been

If the possible disulphide bonds are drawn between the matrix and microfibril to give the unified model (Figure 3), then it is immediately apparent that only a few may be so formed. Presumably the other cystine residues of the matrix and microfibrils are used for the internal crosslinking and stabilization of their respective structures. Thus some parts of the protofibrils are stiffened by covalent crosslinks with the matrix, and others are not. This type of model has also been deduced from a study of the mechanical properties of reduced and methylated fibres34 The proposed structure is attractive in that it incorporates the features of three different mechanical models mentioned above. It is also consistent with current chemical evidence, and the previous apparently contradictory reports from X-ray diffraction observations concerning the regular distribution of cystine in a-keratin are easily explained. ACKNOWLEDGEMENTS The above work was carried out under a Research Studentship sponsored by the Science Research Council to whom the author is indebted for their financial support. The author would also like to acknowledge the P O L Y M E R , 1972, V o l 13, F e b r u a r y

67

X-ray diffraction studies of ~-keratin : G. A. Wilson

considerable help and advice given by Mr H. J. Woods, a n d especially f o r his critical help in p r e p a r i n g manuscript.

the

REFERENCES 1 Simpson, W. S. and Woods, H. J. Nature 1960, 185, 157 2 Dobb, M. G., Fraser, R. D. B. and Macrae, T. P. 3rd Congr. Int. de la Rech. Text. Lainiere, Inst. Text. de France, 1965, Vol 1, p13 3 Bailey, C. J. PhD Thesis, University of Leeds, 1966 4 Spei, M., Heidemann, G. and Zahn, H. Naturwiss. 1968, 7, 346 5 Fraser, R. D. B., Macrae, T. P. and Miller, A. J. Mol. Biol. 1964, 10, 147 6 Dobb, M. G. PhD Thesis, University of Leeds, 1963 7 Cockburn, R. PhD Thesis, University of Leeds, 1954 8 Tyson, C. N. PhD Thesis, University of Leeds, 1969 9 Forrester, M. M. MSc Thesis, University of Leeds, 1958 10 Kulkarni, V. G. PhD Thesis, University of Leeds, 1969 11 Maclaren, J. A. and Sweetman, B. J. Aust. J. Chem. 1966, 19, 2355 12 Block, R. J. 'A Laboratory Manual of Analytical Methods in Protein Chemistry' (Ed. P. Alexander and R. J. Block) Pergamon Press, Oxford, 1960, Vol 2, p 1 13 O'Donnell, I. J. and Thompson, E. O. P. quoted in Adv. Protein Chem. 1965, 20, 232 14 Bailey, C. J., Tyson, C. N. and Woods, H. J. 3rd Congr. Int. de la Rech. Text. Lainiere, Inst. Text. de France, 1965, Vol 1, p21

68

POLYMER, 1972, Vol 13, February

15 Heidemann, G. and Halboth, H. Nature 1967, 213, 71 16 Spei, M., Heidemann, G. and Halboth, H. Nature, 1968, 217, 247 17 Bear, R. S. Adv. Protein Chem. 1952, 7, 69 18 Crick, F. H. C. Aeta Cryst. 1953, 6, 685 19 Fraser, R. D. B., Macrae, T. P. and Miller, A. J. Mol. Biol. 1965, 14, 432 20 Parry, D. A. D. 'Fibrous Proteins' (Ed. W. G. Crewther) Butterworths, Sydney, 1967, p 287 21 Parry, D. A. D. J. Theor. Biol. 1969, 24, 73 22 Parry, D. A. D. and Suzuki, E. Biopolymers 1969, 7, 199 23 Parry, D. A. D. J. Theor. Biol. 1970, 26, 429 24 Astbury, W. T. and Woods, H. J. Phil. Trans. R. Soc. 1933, A232, 333 25 Fraser, R. D. B. and Macrae, T. P. J. Mol. Biol. 1961, 3, 640 26 Onions, W. J., Woods, P. B. and Woods, H. J. Nature 1960, 185, 157 27 Jeffrey, J. M., Sikorski, J. and Woods, H. J. Proc. Int. Wool Text. Conf., Aust. 1956, F, 130 28 Fraser, R. D. B., Macrae, T. P. and Rogers, G. E. Nature 1962, 193, 1052 29 Menefee, E. Text. Res. J. 1968, 38, 1149 30 Skertchly, A. R. B. Nature 1964, 202, 161 31 Corfield, M. C., Fletcher, J. C. and Robson, A. 'Fibrous Proteins' (Ed. W. G. Crewther), Butterworths, Sydney, 1967, p 289 32 Crewther, W. G., Gillespie, J. M., Harrap, B. S., O'Donnell, I. J. and Thompson, E. O. P. 3rd Congr. Int. de la Rech. Text. • Lainiere, Inst. Text. de France, 1965, Vol 1, p 303 33 Crewther, W. G. and Harrap, B. S. J. Biol. Chem. 1969, 242, 4310, 34 Crewther, W. G. Text. Res. J. 1965, 35, 867