Growth of collagen fibril seeds from embryonic tendon: fractured fibril ends nucleate new tip growth

Abstract

Collagen fibrils are the principal tensile element of vertebrate tissues where they occur in the extracellular matrix as spatially organised arrays. A major challenge is to understand how the mechanisms of nucleation, growth and remodelling yield fibrils of tissue-specific diameter and length. Here we have developed a seeding system whereby collagen fibrils were isolated from avian embryonic tendon and added to purified collagen solution, in order to characterise fibril surface nucleation and growth mechanisms. Fragmentation of tendon in liquid nitrogen followed by Dounce homogenisation generated fibril length fragments. Most (>94%) of the fractured ends of fibrils, which show an abrupt square profile, were found to act as nucleation sites for further growth by molecular accretion. The mechanism of this nucleation and growth process was investigated by transmission electron microscopy, atomic force microscopy and scanning transmission electron microscopy mass mapping. Typically, a single growth spur occurred on the N-terminal end of seed fibrils whilst twin spurs frequently formed on the C-terminal end before merging into a single tip projection. The surface nucleation and growth process generated a smoothly tapered tip that achieved maximum diameter when the axial extension reached approximately 13 mum. Lateral growth also occurred along the entire length of all seed fibrils that contained tip projections. The data support a model of collagen fibril growth in which the broken ends of fibrils are nucleation sites for propagation in opposite axial directions. The observed fibril growth behaviour has direct relevance to tendon matrix remodelling and repair processes that might involve rupture of collagen fibrils.

Fig. 1

TEM of fibril seed growth after increasing incubation times in collagen solution. Fibril seeds were released from 13-day chick embryonic metatarsal tendon by crushing at liquid nitrogen temperature and then dispersing in “fibril dispersion” buffer [50 mM Tris–HCl, 50 mM EDTA (ethylenediaminetetraacetic acid), 150 mM NaCl, and 100 mM sucrose (pH 7.4)] in a Dounce homogeniser. The fibril suspension was diluted into a solution of acid-extracted type I collagen (50 μg/ml) at 34 °C in a Na₂HPO₄ (62 mM)/KH₂PO₄ (15 mM) buffer, I 0.2, pH 7.4, which had been set up according to the “warm-start” procedure. The extraction of type I collagen from bovine skin and subsequent purification of a monomeric solution was as described previously. Importantly the telopeptides (extrahelical domains) of the collagen molecule, known to have a critical role in fibril assembly, were preserved intact. A droplet of the seed suspension was placed on a carbon-filmed 200-mesh copper grid and left to adsorb for 1 min, washed with ultrapure water and air-dried. Samples shown were unstained and imaged in a Tecnai-12 transmission electron microscope (FEI, Eindhoven, the Netherlands). (a) Typical fibril length fragment released from 13-day chick embryonic tendon. (b) Projection formed on blunt fibril ends after 30 min and (c and d) after 2 h incubation in collagen solution. (e) and (f) show long tapered projections after 24 h incubation. The arrows indicate the junction of the new tip projections with the seed fibril. (g) Plot of fraction of regrown fibril ends against time. The points are shown fitted to a function of the form f = b(1 − exp(− at)), indicating an average nucleation half-time of 2.1 h and a fraction (6%) of fibril ends that are unable to nucleate fresh growth. (h) Projection length distributions for two incubation times (2  and 24 h). The data are consistent with a near-uniform axial growth rate for the fibril projections. Scale bars, (a)–(d), 0.25 μm; (e) and (f), 2 μm.

Fig. 2

STEM data from seed fibrils after growth. Grid samples of unstained fibrils were prepared as described in the legend to Fig. 1. STEM imaging was on a Tecnai-12 TEM/STEM equipped with a high-angle annular dark-field detector (Fischione, Surrey, UK) with scan and image acquisition controlled by TIA software (FEI). Tobacco mosaic virus (TMV) was used as standard of mass per unit length (131 kDa/nm). (a) Annular dark-field STEM image of part of an unstained fibril tip formed after 24 h growth. The vertical arrow marks the junction of the seed fibril and the new tip projection. (b) Corresponding axial mass distribution measured from the STEM image in (a). The double-arrowed line corresponds to the axial region shown in (a). The plot is typical in showing an initial smooth increase in mass per unit length leading to a plateau region of limited mass per unit length. A plot of maximum mass per unit length versus projection length is shown in (c). The plot shows that a maximum lateral size of projections is attained after a length of ∼ 200 D periods (13.4 μm) has been reached. (d) Comparison of mass per unit length distribution of the initial seed fibrils with that of fibrils after 24 h of growth. (e) Mean values of mass per unit length for 50 seed ends with tip projections. The sites of measurement correspond to either side of the junction (B on seed, C on spur), midpoint along the seed fibril (position A) and the position of maximum mass per unit length along the spur (position D). These values are shown compared with the mean mass-per-unit-length values of the original seed fibrils and those of early fibrils reconstituted from the collagen solution at 200 μg/ml without seeds. On average, the fibril seeds show a 2.6-fold increase in mass per unit length after 24 h of growth. Error bars show SEM values.

Fig. 3

Ultrastructure over the growth junction of seed fibrils. AFM was performed on air-dried samples deposited on freshly cleaved mica using a Veeco Multimode with a Nanoscope IIIa controller operating in intermittent contact mode. (a) Typical AFM image of seed fibrils and tip projections after 2 h of growth. The specimen area is 4 μm × 6 μm and is scaled to 20 nm in height. A fibril plus tip lying between the white arrows in (a) is shown straightened (using ImageJ¹¹) in the upper part of (b). The height scale is 18 nm and is shown displayed with continuous and discontinuous look-up tables. A height plot for a midposition along the fibril axis is shown in the lower part of (b). A dark-field STEM image of an unstained fibril seed–spur junction is shown in (c) and the derived axial mass distribution in (d). The periodic gap–overlap structure is clearly visible in both AFM and STEM and continues in phase over the junction between seed fibril and projection. (e and f) Typical TEM images of seed–spur junctions (vertical arrows) with 1 and 2 spurs, respectively, after negative staining with 2% uranyl acetate. Molecular polarity (N→C), as indicated by the stain pattern, is preserved across the junction as shown. D period = 67 nm.

Fig. 4

Alternative schemes for new tip growth on fibril ends. (a) Possible structural variants of fibril fragments having a primary spur or primary and secondary spurs on each end (N, C) of the fibril. Numbers show the observed frequency of each structural variant. Grey boxes show the two most abundant variants. (b) Model for the progressive addition of collagen molecules to a blunt fibril end whilst preserving the nD overlap between molecules, where n = 1 to 4. (c) Schematic representation of a simple nucleation and propagation growth model consistent with the observed growth forms and the mass per unit length data. The axial mass profiles of the new fibrillar growth (dark grey) will depend on the relative growth rates in solution and along the seed fibril surface. (d) STEM image of an unstained collagen seed fibril after 24 h of growth showing a continuity of microfibrillar substructure extending back from the newly formed tip along the surface of the seed fibril. D period = 67 nm.