Electron Microscopy of Protocollagen Proline Hydroxylase from Chick Embryos

Abstract Protocollagen proline hydroxylase was purified from chick embryos, and it was examined by electron microscopy with negative staining techniques. Enzyme prepared in Helsinki consisted almost entirely of single ring structures. Enzyme prepared in Philadelphia under essentially the same conditions consisted almost entirely of tubular forms in which four rings were stacked together. Repeated freezing and thawing of the Helsinki preparation markedly increased the number of four-ring structures, but it was not possible to establish reproducible conditions for the interconversion of these two major forms of the enzyme. Each ring in the four-ring structure had a diameter of 100 A, a height of 30 to 35 A, and a central hole of about 20 A. The single ring structure was slightly more distended. The molecular weight of each ring was about 200,000, and therefore the molecular weight of the four-ring structure was about 800,000. The inner two rings of the four-ring structure were closer together than the outer two rings. Since the four rings appeared to be identical, the observations suggested that the rings were held together by a combination of heterologous and isologous bonds. No regular structures with more than four rings were seen, and the simplest bonding sequence consistent with these observations was (AB)—(AB)— (BA)—(BA) in which larger structures were not favored, because the formation of the inner B—B bond altered the A surfaces on the inner rings so that they were not equivalent to the A surfaces on the outer rings.


Protocollagen
proline hydroxylase was purified from chick embryos, and it was examined by electron microscopy with negative staining techniques.
Enzyme prepared in Helsinki consisted almost entirely of single ring structures.Enzyme prepared in Philadelphia under essentially the same conditions consisted almost entirely of tubular forms in which four rings were stacked together.
Repeated freezing and thawing of the Helsinki preparation markedly increased the number of four-ring structures, but it was not possible to establish reproducible conditions for the interconversion of these two major forms of the enzyme.
Each ring in the four-ring structure had a diameter of 100 A, a height of 30 to 35 A, and a central hole of about 20 A. The single ring structure was slightly more distended.
The molecular weight of each ring was about 200,000, and therefore the molecular weight of the four-ring structure was about 800,000.
The inner two rings of the four-ring structure were closer together than the outer two rings.
Since the four rings appeared to be identical, the observations suggested that the rings were held together by a combination of heterologous and isologous bonds.
No regular structures with more than four rings were seen, and the simplest bonding sequence consistent with these observations was (AB)-(AB)-(BA)- (BA) in which larger structures were not favored, because the formation of the inner B-B bond altered the A surfaces on the inner rings so that they were not equivalent to the A surfaces on the outer rings.
The hydroxyproline,in collagen is synthesized by the hydroxylation of proline in a proline-rich and lysine-rich polypeptide precursor of collagen called protocollagen (for reviews, see References 1 and 2).Protocollagen proline hydroxylase has been identified in connective tissue from several sources, and has been shown to require 02, Fez+, a-ketoglutarate, and, perhaps, ascorbate.
The enzyme does not hydroxylate free proline, tripeptides such as Gly-Pro-Pro, or the polymer poly-L-proline, but it hydroxylates polymers of the structure (Gly-X-Pro),, in which X is proline, alanine, or a variety of other amino acids but not glycine (1).
Recent,ly, protocollagen proline hydroxylase has been purified extensively from chick embryos (3,4) and from fetal rat skin (5).The enzyme from chick embryos was obtained in a form which was homogenous by ultracentrifugat,ion and which was over 90% pure by disc electrophoresis (4).This enzyme had a s20,W of 6.7 S, which suggested a molecular weight of about 150,000 (4).In the present paper we describe electron microscopic studies of the enzyme from chick embryos.
The results indicate that the enzyme consists of ring structures of about 30 A by 100 A4 which are seen either as isolated rings or in tubular forms containing two or four rings.

MATERIALS
AND METHODS E'nzyme Pur$ccation-Protocollagen proline hydroxylase was prepared from 12-day-old chick embryos with the procedures recently described by Kivirikko et al. (3,4).About 400 embryos from white Leghorn chickens were used as a source of enzyme, and the purification procedure consisted of ammonium sulfate fractionation, fractionation with calcium phosphate gel, and chromatography on a DEAE-cellulose column (Whatman, DE-23).
The column was equilibrated with 0.2 M NaCl, 0.2 M glycine, and 0.01 M Tris-HCl pH 7.0 at 4", and the enzyme eluted with 0.30 M NaCl was concentrated by vacuum dialysis or by ultrafiltration.
The enzyme was then chromatographed on a column, 2.5 X 95 cm, of Syc agarose (Bio-Gel A, 1.5 m, 100 to 200 mesh; Bio-Rad) which was equilibrated and eluted with 0.01 M Tris-HCl, pH 7.8, at 4", containing either 0.2 M NaCl and 0.2 M glycine, or 0.05 M KCl, 0.1 M NaCl, and 0.1 M glycine.
The flow rate of the column was about 80 ml per hour and IO-ml fractions were collected.
Disc electrophoresis was carried out in a 6.5y0 polyacrylamide gel with Tris-glycine buffer (4).The pH of the upper buffer was 9, the pH of the lower buffer was 8.2, and the gel was stained with Amido black.Enzymatic Assay-The enzyme was assayed in an incubation system containing 125 pg per ml of the polytripeptide (Pro-Pro-Gly)~ ( 6), 20 to 250 pg per ml of enzyme protein, 0.5 m&f acketoglutarate, 0.05 mu FeS04, 2 mu ascorbic acid, 0.2 mg per ml of catalase (Calbiochem), 0.1 mM dithiothreitol (Eastman Organic Chemicals), 2 mg per ml of bovine serum albumin, and 50 mM Tris-HCl buffer, adjusted to pH 7.8 at 25", in a final volume of 4 ml (3)(4)(5).
Samples were incubated at 37" with shaking for 1 hour, and the reaction was stopped by the addition of 4 ml of concentrated HCl.
The samples were hydrolyzed overnight in sealed tubes at 120" and the hydrolysates were evaporated to dryness in a rotary evaporator.
The residues were dissolved in 4.0 ml of distilled water, and the hydroxyproline content was assayed by a specific chemical procedure (7) in 3.5-ml aliquots.
As indicated elsewhere (4), the standard assays were not carried out with saturating concentrations of substrate because of the difficulty and expense in obtaining appropriate polypeptide substrates.
All values for enzyme units and specific activities referred to here were based on assays carried out under the conditions described above.One unit of enzymatic activity was defined (3,4) as the number of micrograms of hydroxyproline synthesized in 1 hour under the standard assay conditions.The amount of hydroxyproline synthesized in 1 hour was about 2-fold greater when saturating concentrations of substrate were used (4).Because the K, value for the (Pro-Pro-Gly), substrate was somewhat higher than for polymer preparations of (Pro-Gly-Pro), with an average molecular weight of about 3000 (4,8), the values for units of enzyme observed here were corrected by a factor of 1.25 in order to make the values directly comparable to those reported previously.
Techniques for Electron Microscopy-Protocollagen proline hydroxylase was dissolved in 0.01 M Tris-HCl, pH 7.8, containing either 0.2 M NaCl and 0.2 M glycine or 0.05 M KC1 and 0.1 M glycine, and it was examined by negative staining (9).One drop of the solution was applied to a carbon film on an electron microscopic grid.The carbon film was prepared by evaporating carbon onto Formvar-coated grids and the Formvar was removed by placing the coated grids on a filter paper soaked in chloroform.After application of the enzyme solution to the carbon film on the grid, the edge of the grid was blotted with filter paper.The thin film thus produced was allowed to evaporate for a few seconds.Negative staining was then performed by adding 1 drop of 1% uranyl acetate (E.Merck, Darmstadt, Germany) at pH 4.8, or 1 drop of 4% sodium silicotungstate (Taab Laboratories, Reading, England) at pH 7. The excess fluid was blotted with filter paper.
In some instances the enzyme solution was mixed with an equal volume of 4% sodium silicotungstate prior to application to the grid.
The grids were examined in a Hitachi HU-llC-1 electron microscope and Kodak contrast projector slide plates were exposed to the image at a magnification of 100,000.The accelerating voltage was either 75 or 100 kv.

Preparation of Enzyme-All
of the studies reported here were carried out with protocollagen proline hydroxylase prepared from chick embryos by the most recent procedure of Kivirikko et al. (3,4).Most of the work was done with enzyme prepared in Philadelphia, but, because gel filtration indicated that the molecular weight of the preparations was greater than that reported previously (4), the electron microscopy was repeated with enzyme prepared in Helsinki.The enzyme was purified through the DEAE-cellulose column and then it was chromatographed on an 87, agarose column as described in the text.Elution peaks of other substances in the same column: I, blue dextran (mol wt --2,OOO,OOO) ; 6, IgM globulin (mol wt ~l,OOO,OOO) ; 3, peak fraction for protocollagen proline hydroxylase prepared in Helsinki; 4, IgG globulin (mol wt ~150,000) ; 5, albumin (mol wt -65,000); 6, cytochrome c (mol wt -13,000).

Elution of protein as indicated by absorbance at 230 rnp, O-O
; elution pattern of protocollagen proline hydroxylase, O-0.In preparing the enzyme in Philadelphia, the results with the initial two steps of ammonium sulfate and calcium phosphate gel fractionation were about the same as those obtained in Helsinki (3, 4), but differences were found during chromatography of the enzyme on the DEAE-cellulose and on the 8% agarose columns.
In Philadelphia, the recovery of enzyme The letter Note that the distance between the inner two rings is less t'han the M at the top of the picture indicates a short thread-like structure distance between the outer and the inner rings.X 550,000.
from the DEAE-cellulose column was better than in Helsinki, separate preparations.However, the best enzyme preparations and 89 to 95% of the enzyme was recovered in the 0.3 M and obtained after these two steps had a specific activity of only 0.4 M NaCl fractions with three separate preparations.
The re-about one-third of that obtained in Helsinki (3,4).These differcovery of enzyme in the agarose column was also better in Phila-ences were consistent even though the procedures used were delphia, and 90 to 94% of the enzyme was recovered with three apparently identical.
FIG. 4. Negative staining with 4'$& sodium silicotungstate of structure.X 1,000,000.Inset, left half shows a top view of a enzyme prepared in Philadelphia.
The protein concentration was four-ring structure with an hexagonal outline.
Right half shows a 12Opg per ml.T, top and side views of four-ring structures; D, rotational print of the same image assuming a B-fold symmetry.side views of two-ring structures; M, top view of a single ring x 1,800,000.
The enzyme prepared in Philadelphia consistently eluted near the void volume of the 8% agarose column (Fig. I), and the ratio of the elution volume to the void volume (V,/V,) was 1.10 to 1.20.The enzyme prepared in Helsinki consistently eluted with a V,/Vo of 1.40.Identical V,/Vo values were obtained when the columns in both laboratories were standardized with blue dext'ran, IgM globulin, serum albumin, and cytochrome c (Fig. 1).
below to indicate that the rings are top views and the tetrad forms are side views of the same molecule, and we conclude that most of the molecules in this preparation are tubules composed of four-ring structures stacked together.
The presence of four rings in the tetrad forms is clearly shown in Fig. 3 (inset).The same picture also shows that within the tubule the distance between an outer ring and an inner ring is greater than the distance between the two inner rings.Disc electrophoresis in Tris-glycine buffer of the enzyme prepared in Helsinki gave one minor and one major band (4).The major band was well within the gel, it contained enzymatic activity, and it accounted for over 90% of the total protein (4).Under the same conditions the enzyme prepared in Philadelphia (Fraction 23 of the chromatogram shown in Fig. 1) gave four minor bands and one major band.The major band accounted for most of the total protein, but it was only 1 or 2 mm from the stacking gel.Electron microscopy of the enzyme prepared in Philadelphia showed that essentially all of the material consisted of regular enzyme structures (see below), and therefore the results indicated that the major band obtained with disc electrophoresis was t,he enzyme.
Most of the top views show the presence of a central hole with a diameter of 20 to 25 A, and under favorable staining conditions a channel of this width is occasionally seen in side views of the molecule (arrows in Fig. 3).
Although most of the molecules apparently consisted of four rings, some top views were of lower contrast than others (Figs. 3 and 4) and some side views indicated only one or two rings (Fig. 4).These variations are probably explained by the enzyme existing in the form of single-ring, two-ring, and four-ring structures.
No structures which could be interpreted as consisting of three rings or of more than four rings were observed.
General Features Reveakd by Electron Microscopy-The initial electron microscopy was carried out with Fraction 23 of the chromatogram shown in Fig. 1, which had a specific activity of 86 units per mg of protein.
After negative staining with either uranyl acetate or sodium silicotungstate, rings about 100 A in diameter and tetrad structures of about 140 A in length were observed (Fig. 2).The same structures were found both when a dilute solution containing 40 pg of protein per ml was examined and after the solut.ionwas concentrated five times by vacuum dialysis.
The same structures are seen at higher magnification in Fig. 3.We interpret these figures as well as those shown The single-, two-, and four-ring structures seen in the Philadelphia preparations were also found in enzyme preparations made in Helsinki (Fig. 5, A and B).However, the characteristic feature of the Helsinki preparations was a high proportion of top views of ring structures which were probably single rings.The inner and outer diameters of the single rings were consistently about 20 A greater than the diameters of the four-ring structures.
The single rings may have appeared larger either because they were larger in solution or because the rings were partially distorted by the stain or by the hydrophobic surface of the carbon film.In addition, we frequently observed short thread-like structures of 30 to 35 A by 100 to 120 A which probably represented side views of single rings (Fig. 5B).The B. R. Olsen, X. A. Jimenex, K. I. Kivirikko FIG. 1. Chromatographyof protocollagen proline hydroxylase prepared in Philadelphia.The enzyme was purified through the DEAE-cellulose column and then it was chromatographed on an 87, agarose column as described in the text.Elution peaks of other substances in the same column: I, blue dextran (mol wt --2,OOO,OOO) ; 6, IgM globulin (mol wt ~l,OOO,OOO) ; 3, peak fraction

FIG. 2 .
FIG. 2. Negative staining with 47, sodium silicotungstate of the enzyme recovered in tube 23 of the chromatogram shown in Fig. 1.The concentration of protein was 120 pg per ml and the pH of the solution was 7.8.A large number of tetrad side views and a few circular top views are seen.(See text for definition of "side views" and "top views.")X 240,000.

B
FIG. 3. Examination of the enzyme preparation shown in Fig. 2 of about 100 A in length which could represent a single ring in side at a higher magnification.The arrows indicate tetrads in which view.The open arrow indicates a top view of a structure with a the central channel is seen in side view.Ring structures of low triangular shape.Inset, side views of tetrads in which the concontrast which probably represent single rings (see t.ext) are indi-trast of the image was increased by rephotographing the prints.cated by the letter M in the lower part of the picture.The letter Note that the distance between the inner two rings is less t'han the M at the top of the picture indicates a short thread-like structure distance between the outer and the inner rings.X 550,000.