Structural studies of the outer envelope of Chlamydia trachomatis by electron microscopy

Structural studies of the outer envelope of Chlamydia trachomatis by electron microscopy

J. Mol. Biol. (1982) 161, 579-590 Structural Studies of the Outer Envelope of Chlamydia trachomatis by Electron Microscopy ,Jris-.rr. (:HAS(: The In...

11MB Sizes 0 Downloads 3 Views

J. Mol. Biol.

(1982) 161, 579-590

Structural Studies of the Outer Envelope of Chlamydia trachomatis by Electron Microscopy ,Jris-.rr. (:HAS(: The Institute

of Biophysics

of Academia

KEVIN LEOWU),

Postfach

t’ekin,g.

%HAM:

Drpartmrnjt Beijing

(l&wived

Institute 22 April

China

T,\~,nlos AKAI), TOW PIW

European Molecular Biology Laboratovy 10.2209, D-6900 Heidelberg, Federal Rcpublir YOl--XI'S

Electron

Rinicz~,

of Germany

ASI) IAILHI'.4 %HAN(: of Hiochemistry

of Ophthalmology,

J’ekir~g,

19X2, and in revised form

China

22
observations have been made on the outer cell rnvrlopes of the agent of trachoma. It has been shown that they contain a hexagonal lattice with a unit cell of about 17.5 nm. (‘omputer processing has been carried out on the hexagonal array and on ring-like particles occurring in of these rings can be related to the fragmented envelopes. The structurta distribution of material in a three-dimensional model obtained for the hexagonal lattice. This model has been compared with biochemical data for the cell envelopes. C’hlumydia

microscopic

trachomatis,

1. Introduction Chlamydiue are obligate intracellular bacteria, which cause a number of diseases in man. They are prokaryotic organisms but their life cycle depends upon development in host cells (for a review, see Becker, 1978). Two different developmental stages can be distinguished, the infectious elementary body, which is less than 0.5 pm in diameter and the non-infectious reticulate body with a diameter of about 0.6 to 14p.m (Higash, 1965). This difference in behaviour is thought to be accompanied by a change in the Chlamydia cell envelope (Matsumoto & Manire, 1970) and thus study of this structure by electron microscopy may lead to a better understanding of the process of infection by these organisms. The cell wall structure of Chlamydiae is similar to that of gram-negative bacteria (Sleytr, 1978) but has been less well characterized. Our previous studies on the envelope of elementary bodies of Chlamydia trachomatis have shown that it consists of an outer cell envelope, also called “cell wall”, and an inner, cytoplasmic membrane. In the outer cell envelope, a regular hexagonal lattice of material is observable by negative staining (Zhang et al., 1980). 5h‘imilar observations have also 0022%2836/82/32057Crl2503.00/0

579

0 1982 Academic Press Inc. (London)

IAA

5x0

.J.-,JI:

(‘HAS(:

K’7’ .-I/,

made for ChZamydia psittaci (Matsumoto 02 Manire, 1970). This regular arrangement of protein subunits can be considered as the equivalent of the bacterial S-lager. In this paper. we describe the ultrastructure of the out,er cell envelope of (‘. tmchomatis. which has been studied in more detail by three-dimensional reconstruction of the hexagonal array. Finally? we will compare our results with those of biochemical investigations. been

2. Materials The TE.55 strain of C. tmchomntis. cultured in these studies. (‘rll culturing and purification

and Methods by Tang’s method (Tang ut ol.. 1957). \vas used \vere carried out as before (Zhang pi 01.. I980).

(a) Z’r~~tm~,
(b) Eluctrotl

nticroacopy

Negatively stained specimens were prepared as follows. About 5 ~1 of treated Chlamydia solution was applied to hydrophilic carbon-coated 400-mesh grids freshly glow-discharged in air. After 1 min for adsorption a drop of 1% phosphotungstic acid (pH 7.2) staining solution was added. Excess liquid was removed with filter paper and the grids dried in air. Micrographs were taken in a Philips EM 400 electron microscope operating at 80 kV and 28,000 x magnification using th
ESVELOPE

STRI’CTURE

OP C. trachomalis

5X1

(‘omput,rr noise filtered images were made by transforming the peak amplitudes and phases of the Fourier map. The individual ring-like particles in zero-tilt images wer(x scanned under the same conditions and were then subjected to rotational averaging using the Semper program package (Saxton et al., 1979). Circular sections were extracted from the ring and an autocorrelation was then calculated in one dimension along the sections. The resulting angular correlation function shows a peak at each angle at which a good match is found. so that angular symmetry shows up as a set of peaks spaced regularly across the function. Thcs rotational average of each ring was computer calculated by superposing the image. rotjated on itself the appropriat,e number of times hy the correlation angle.

(d) Three-dimension&

reconstruction

Each set) of data at different angles was separatelv refinecl to a common phase origin in reciprocal space. The refined phases fall within 30” for 13 images from +35” to -50 Refinement of phases and amplitude scaling was carried out. between reflections closrr than l/50 nm in reciprocal space. This was first done using the Z-sided plane group pl (Holser. 19.58) and then, since the phase bchaviour corresponds closely to that of p6 symmetry. the refinement was also carried out in space group ~6. Data from the combined titled images were used to plot variation of amplitude and phase in the Z* direction. These curves (see Fig. 4) were fitted by hand and interpolated at intervals of l/30 nm. A 3-dimensional Fourier synthesis was then calculated in either space group pl or ~6. No significant improvement was obtained by combining all the data into one map.

3. Results (a) Electron

microscopy

The negatively stained preparations show that for the phosphotungstic acidstained envelopes of C. trachomatis digested by sodium desoxycholate, a clear hexagonal lattice structure is present (Fig. 1). It not only exists in small elementary bodies (Fig. l(a)) but also in large reticulate bodies, and during binary fission (Fig. l(b)) for which optical diffraction reveals double layers indicating that the envelope has not been broken. This hexagonal array is also present in uranyl acetate-stained envelopes but it is neither as clear, nor was the resolution as good in optical diffraction as was obtained for. phosphotungstic acid-stained envelopes, Apart from the image in Figure l(b), all micrographs shown are for small elementary bodies. Digested envelopes treated by sonication and then shadow-cast with platinum show that the single layer structure is about 11 nm in thickness using tobacco mosaic virus as internal standard. In partially broken envelopes, it can be seen that the outside is relatively smooth, whereas the inner side facing the cytoplasmic membrane shows a clear hexagonal lattice (Fig. Z).‘Similar observations have been made for C’hlamydia psittaci by Matsunioto & Manire (1970). Occasionally stick-like protuberances about 10 nm in diameter and 3.5 nm in length can be seen on the outside of C. trachomatis (Fig. l(a), arrow and Fig. 2). Figure 3(a) shows an envelope that was broken into a single layer by ultrasonication and stained with phosphotungstic acid. The optical diffraction of this kind of single layer (Fig. 3(b)) shows the unit cell of the hexagonal array to be

b)

P’I(:. 1 Shadowrast envelopes of P. tmchon~ntis mixed with tobacco mosaic virus. Purified C’J~kvn~d~u~ wtw digested for 2 h in 1’3, sodium desoxycholate follorrd hy slight ultrasonication. Note that tht outsidr is smooth and the hexagonal array is visible only on the inner side. Magnification, 65,0(K) x

FIG:. 3. (a) Segmented envelopes of C. trnchomatis digested by 19/o sodium desoxycholate for 2 h and sonicated. Segativrly stained with phosphotungstic acid. Note the single laker, which shows both a regular hexagonal array and a disordrred region containing individual ring-llkr structures (arrowed). Magnification, 13,000 x (11) The optical diffraction pattern of a single layer. Note the G-fold symmetry of the intensitirs

ENVELOPE

STRUCTURE

OF

5x5

C. trachomatis

about 175 nm. The pattern shows good 6-fold symmetry for the peak intensities. Evidence from calculated phases and filtered images show that it is indeed ~6. Individual ring-like structures can be seen on some negatively stained specimens (Fig. 3(a), arrows). Their sizes and shapes vary slightly. Similar structures have also been found in elementary bodies of C. psittaci (Matsumoto, 1973).

(b) Three-dimensional

of the hmxqonal

structure

array

Two tilt sets. each of which included 13 images of a single crystalline area, were separately used for calculating the three-dimensional structure. Examples of variations of amplitudes and phases for one tilt set in the Z* direction are shown in I

0

04

O-2

0

04

0.2

0.2

3.1

----J

5400

360°

nm-’

FIG. 4. Sample curves of the variations in amplitudes (+ ) and phases (0) for 4 reflections. the ordinate are given only for phases. The units on the abscissa am for Z*.

[Inits

01)

.J.-.IL‘(‘H.AS(:

67’

.I/,

Figure 4. Density maps were contourrtd at a level whcrt~ a clear edge to the struct,ure was apparent in the direction perpendicular to the plane of the cell wall. Essentially the same result,s \vere obtained for two three-dimensional p6 maps, which were calculated from the t)wo differt>nt. tilt sets. On one side of the hexagonal layer are arranged six subunits of diameter about 3.5 to 4.0 nm around an 8-nm deep depression in the surface, the other side being relatively smooth. These subunits are related as dimers to subunits in the adjacent. cells. There also appears to br connecting density between three dimers across the Xfold symmetry axis of the structure. A solid wooden model (Fig. 7) shoL+ing one hexagonal prism, making up the unit, cell. was constructed for one of t’hc two tilt, s&s. for whiczh the filtered zero-tilt image is also shown in Figure 6(a). Figure 7(a) shows the structure viewed as from the inside of the outer envelope \vith a depression of about IO nm diameter and 8 ntn depth. Figure i(h) shows tht, opposite view, from the outside. The thickness from the model was estimated t,o be about 13 nm (Fig. 7((a) and (d)). which is consistent with t,he estimated thickness from shadowing, the lower value in the latter case (1 1 nm) perhaps resulting from shrinkage during drying. For comparison with t’he filtered image (Fig. 6(a)), a projection of this structure down the c dire&on is also shown (Pig. 6(b)). Ti 11sis &fold averaged and contains information from all images in the tilt set,. so that a much clearer picture of the pro,jected dfmsity is obtitined.

Figure 3(a) shows individual rings (indicated by arrows) after the hexagonal arrays have been partially destroyed by detergent treatment and sonication. Computer rotational analysis of the rings (Frank et al., 1978) shows that in most cases B-fold symmetry was detectable (Fig. 5). In Figure 6(c) and (d) we show a single ring subjected to rotational averaging assuming &fold symtnetry. Six clear domains of density are apparent, which correspond to the marked part of the model (Fig. 7(a)). The average outside diameter of the rings is about 15 nm.

ENVELOPE

STRI.(‘TllRE

OF C. trnchomatis

5x7

FIG. 6. (a) Computer noise filtered image for the zero-tilt micrograph of the tilt set used for the 3dimensional reconstruction shown below. Regions corresponding to high protein density are white and conrervely those corresponding to high stain density are dark. (b) Computer 6-fold symmetrized image obtained by projecting the 3-dimensional map of F’lg. 7 down the c direction. In (a) and more clearly in (b) can be seen high projected protein density in the P-fold positions. lowr~r protein den&v at the 3.fold positions and the lowest protein density in the B-fold positions at the middle of the ring. (c) Computer-processed images of 4 individual rmgs (taken from Fig. 3(a)). (d) A contour/density plot, of the ring in the upper-left of the image in (c). This shows. more clearly. features of lower density at the centre of the annulus and the 6 regions of higher density surrounding it. (a), (b) and (c) art’ thr samcamagnification.

(d) Protein

com$o&tion

In order to have some biochemical data to compare with this model, we have made a preliminary sodium dodecyl sulphate/polyacrylamide gel electrophoretic analysis of the cell wall. Purified Chlamydia were dissolved in 1% sodium dodecyl sulphate by heating at 95°C for five minutes, and then run in a 12.5% sodium dodecyl sulphate/polyacrylamide slab gel. The major band corresponded to a

588

.J.-JlI

(‘HAS{:

EIT

.IL

FIG:. 7. Views of a model for one hexagonal unit cell. (a) A view from the side facing the cytoplasmic membrane. The density seen in an individual ring (Fig. 6(d)) is contained within the broken circle. Segments of broken circles indicate the positions of adjacent unit cells within the hexagonal lattice. The 6 subunit domains, which make up the annulus around the central depression. are emphasized by shading. These are related by a dyad axis t.o equivalent subunits in adjacent unit cells. There is also some connecting density at the S-fold positions. (I)) Opposite vie\v. from the smooth side of the hexagonal layer. (c) Side-virw of the model from outside. (d) Side-view of a selection through the model showing the central depression. In (c) and (d). the density domain at the Z-fold axis, seen shaded in the vies- of (a). is outlined by a broken lim.

molecular weight of Similar results have found a major band more than 60% of

40,000, and two other strong bands to about 50,000 and 35,000. been reported for C. trachomatis by Caldwell et al. (1981), who corresponding to a molecular weight of 39,500, accounting for the total protein of the outer envelope. Tamura et al. (1974),

ESVE:LOPE:

STHI’(‘TC’K.&

OF C’. TH.-1PHOM.-l7'/S

5x9

Hatch rt ul. (1981) and Salari & Ward (1981) also reported the presence of a cell-wall protein in t,he 40.000 molecular weight range in C. trachomatis and C. psittaci.

4. Discussion A hexagonal protein array with unit cell of about, 17,s nm exists in t,he outer &I side of this layer facing t,hc of C. trachomatis. with the “rough” envelope cytoplasmic membrane. It is made up of annular structures, each composed of six subunit domains surrounding a cent’ral depression 10 nm in diameter and 8 nm deep. These are packed so that subunits in adjacent rings are in contact’ across the P-fold symmetry axes. The other side of the outer envelope is relatively smoot,h. IVe have observed this hexagonal lattice structure in all stages of Chlamydia development,. in both small elementary bodies and in large reticulate bodies, and during binary fission (Fig. 1). The large reticulate bodies are, however. mortl sensitive to detergent digestion than the small ones. The ring-like structures shown in Figure 3(a) have the same shape. projected density distribution and approximate size as the central part of the threedimensional reconstruction. They appear to be remnants of the hexagonal array after it has been destroyed by chemical or physical treatment. Thus, protein protein interactions within the ring are evidently stronger than those between dimers. The isolated rings tend to collapse inwards. so that. their diamet,er is reduced by up t,o 15(?; (from 175 to about 1.5 nm). (:el electrophoresis data obtained in this study and by other workers indicate that a polppeptide of about 40,000 molecular weight is the major structural protein of the outer envelope. Calculations made on our model suggest that the volume of each of t,hc unit,s t,hat make up one-sixth of the ring-like structures and that occur as dimers. would correspond to about this molecular weight and can be identified bentatively with t,his major protein. There remains, however, a large part of the structural density seen in the model that is unaccounted for, both at t.he S-fold positions and at the base of the central depression. To assign this density, a more exact’ knowledge of the biochemistry and subunit stoichiometry of the cell wall will be required. The

authors

C. trachomatis.

polpacrylamide

thank Miss Juan-juan Zuo for the cultivation and purification of Alan Darcy and John Dixon for assistance with sodium dodecyl sulphate] gel electrophoresis and Dr J. Lepault for the shadow-cast preparations.

RE FERENCER Becker, Y. (1978). Microhiol. Rev. 43, 274-306. Caldwell. H. D.. Kromhout, J. & Schachter. J. (1981). Irlfect. Immun. 31, 1161-1176. Frank, J.. Goldfarb, W., Eisenberg, D. & Baker, T. S. (1978). Ultramicroscopy, 3, 283-290. Hatch. T. I’., Vance, D. W. & AI-hossainp, E. (1981). J. Bacterial. 146, 426-429. Higash, N. (196.5) Exp. Mol. Pathot. 4, 24-39. Holser. W. T. (1958). 2. Kristallogr. 111, 266-281. Leonard. K., Wingfield, P.. Arad, T. & Weiss. H. (1981). J. Mol. RioZ. 149. 259-274. &iatsumoto. A. (1973). J. Bucteriol. 116, 1355-1363. Matsumoto. A. Br Manire, G. 1’. (1970). J. Bactcriol. 104 133221337.

.I.-.JI’ (‘HASCI

590

F7’ > .-I/,

Salad, S. H. B Ward. RI. E. (1981 ), ./. (~PI/. Microhiol. 123, 19-205. Saxton. W. O., Pitt. T. J. & Homer. JI. (1979). ~‘ltmrr~icmscopy. 4. 343-354. s1tytr. 11. R. (1978). Id. RP?‘. (‘@Id. 53. 1~64. Tamura, :I.. Tanaka. .A. & Manire, (:. I’. (l!Uic). ./. /hrcterio/. 118. IS!) 1-U. Tang. F. P.. (‘hang. H. I,., Huang. Y. T. & W~II~. K. (‘. (195-i). (‘hi,/. Med. ./. 75. ti!) 147. Zhatlg. Y. ?(.. Mcrrg. S. .\I.. Zhatrg. I,. H.. Srt. H. k I.i. R. I). (1980). AScic,olirrSitticrc. 23. 120X-1217.