Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by GLASGOW UNIVERSITY LIBRARY on 04/26/13 For personal use only.

The response of cell walls of Bacillus subtilis to metals and to electron-microscopicstains T . J. BEVERIDGE Depcirit~tetriof Btrc.ieriology trrlcl I t ~ ~ t ? ~ r r t ~ o lU og t ~yi.~ ~ e r sofi i yWes/c,rtr Otticrrio, Lotrclotr, Otri. , Cntlciclct

Accepted October 24, 1977 B E V E R I D GTE., J . 1978. The response of cell walls of Bocil1rr.s .srrbiili.s to metals and to electronmicroscopic stains. Can. J. Microbiol. 24: 89- 104. Purified cell walls of Btrcillrrs .srrbrili.s were subjected to solutions of 40 independent metals and the metal uptake, the electron-scattering power of thin sections, and the type of staining response evaluated. This was repeated for six typical electron-microscopic stains (uranyl acetate, uranyl magnesi~lm acetate, osmium tetroxide, 0s-meth, osmium-dimethylethylenediamine, and ruthenium red) and one new staining reagent (a potassium platinum chloride - dimethylsulfoxide complex) whose specificity is for amine functions. The reaction of select metals can be specific in terms of both uptake and staining response. Of the metals studied most transition elements had a highaffinity for the wall fabricand some(i.e., Sc 111, most lanthanides, U IV, Zr IV, HfIV, Fe 111, Pd 11, Ru 111, and In 111) may be suitable as contrasting agents for electron microscopy. Furthermore, when the thickness of metal-reacted walls was compared to freeze-eLch and ~lltracryotomydata, statistical-dimensional differences were commonly seen, which indicates that wall ~ ~ l t ~ . a s t ~ . ucan c t ~be~profoundly re affected by the typeof metaland(or) stainingreagent. BEVERIDGE, T. J. 1978. The response of cell walls of Bocillrrs .srrl~iilisto metals and to electronmicroscopic stains. Can. J. Microbiol. 24: 89- 104. Des parois cellulaires purifiees de Bocillrrs sr~biilisfurent soumises a d e s solutions de40 mCtaux independants, puis evaluees par rapport I'absorption des metaux, la capacitk de diffraction d e s electrons en coupes minces et le type de coloration obtenue. Ceci a e t e repet6 pour six colorants propres a la microscopie electronique (I'acetate d'uranyle, I'acCtate d'uranyle et de magnesium, le tetroxideosmique, 1'0s-meth, I'osmium-dimethylethylknediamineetle rougede ruthenium) e t un nouveau reactifcolorant (uncomplexechloruredeplatineet depotassium- dimethylsulfoxide) specifique pour les fonctions amines. L a reaction des metaux choisis peut i t r e specifique e n termes a la fois d'absorption et de capacite de coloration. Des metaux Ctudies, la plupart des elements de transition ont une haute affinite pour le materiel parietal c'est-a-dire, S c 111, la plupart des lanthanides, U IV, Zr IV, Hf IV, Fe 111, Pd 11, Ru 111, and In 111); ils peuvent etre des agents contrastants convenables pour la microscopie electronique. De plus, lorsque I'epaisseur des parois ayant rCagi aux metaux est comparke aux donnees par ultracryotomie e t decapage afroid, il existe des differences statistiques de dimensions qui indiquent que I'ultrastmcture parietale peut-etre profondement affectee par le type de metal et (ou) le reactif de coloration. [Traduit par le journal]

Introduction Since most biological material possesses little electron-scattering ability, electron microscopists have been forced to depend on heavy metal impregnation of the specimen before obtaining thin-section data. In fact, all techniques for visualizing the form and structure of biological material by electron microscopy, whether by replica or negative or positive staining, rely on the use of heavy metals in one form or another. The chemistry of the cell wall ofBcrcill~rss~rbtilis has been studied in detail (5, 10, 19) and, in general, its staining response to the common electron-microscopic stains is typical of the Gram-positive bacterial wall. Bncill~rs s~rbtilis is ubiquitous throughout nature and its wall is the most durable of its organic constituents. For this reason we have used the wall as a model system to study the physicochemical interaction of its component parts

with the metals of the environment. Recently, we have demonstrated its ability to trap and retain substantial amounts of metal ions when suspended in aqueous solutions and that these entrapment sites may be select (1). This report expands these findings to include many more metals, and their potential as electron-microscopic stains is evaluated. Since the electron-microscopic image produced by the more common staining reagents (i.e. uranyl acetate, osmium tetroxide, ruthenium red, etc.) must be a response of the wall to their aqueous ionic form and charge, the i ~ a g e sproduced by several stains have been also compared and evaluated in parallel with the data derived from the metal salts. Materials and Methods Cell WnII Prepcrrcriiotr Cell walls of Boci1lrr.s srrbiilis Marburg (University of Western

Ontario collection No. 1032) were prepared from an exponen-

90

C A N . J . M I C K O R I101. VOL.. 24. 1978

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by GLASGOW UNIVERSITY LIBRARY on 04/26/13 For personal use only.

tially growing culture a s previously described ( I ) . T h e purity of the w;ill fragments was ascertained by electron microscopy (negative stains and thin sections) and by testing for reduced n~cotinamideadenine dinucleotide oxidase as a marker for pliisma membrane enzymes (13).

other than the metal salt reactants were used during the processing for electron microscopy.

//I(,Coil W(r1l.v T h e metal uptake proceduresfor the quantitativeexperiments were the same a s for the electron microscopy. Where applicable Ac/oc~orr.sMrtol Sol1 Solrrtiotrs U.setlJi~rE.rpt~~.ir?~etritriiotr the Millipore membrane-filtersystem was used and isoutlined in The following salts were used for experimentation: LiCI, ref. I. In both cases the walls were extensively washed with 5 NaCI. KCI, RbCI, SrC1,.6H,O, BaCIZ.2H,0, MnCI2.4H?O, volumes of 10 ml DDW. FeCI,.hH,O. CoCI,, Ni(NO,),.hH,O, CuC1,,2H,O, Most metzlls were detected by atomic abso~.ptionanalysis using a Perkin-Elmer model 403 atomic absorption unit (PerAgNO;. ZnCI,. AI,(SO,),- 18H,O. HgCI,, TIC,H,O,, UO,(C,H,O,),.?H,O, L:iCI,.hH,O. La(NO,),.6H2O, kin-Elmer Canada Ltd.) in either the flame or graphite furnace l'tCI,..iH,O. Pb(NO,),. Bi(NO,),.SH,O (all reagent grade: mode by previously described techniques ( I ) but the transition elements I and indium, thallium, and bismuth weredetected by Fisher Scientific Co.): CaCIZ, MgCI,.hH,O (Sigma Chemical Co.): OsO, (BDH Chemicals): Ce(NO,),.hH,O (Electronic the X-ray fluorescence technique using a Philips P W I450 Space Products Inc.. Los Angeles. CA): ScCI,, Atomic sequential X-ray spectrophotometer. For X-ray diffraction analyses small samples of the plastic S m , ( S 0 , ) , . 8 H z 0 . %rOC12.8H,0. H f O C I Z . 8 H , 0 . VOSO,. embeddings used for electron microscopy were mounted on I'dSO,. RuCI,,H,O. AuCI,.2H,O. In,(S0,).3,9HZ0. PrCI,, glass fibers and each suspended in the center of a Gandolfi CdCI,. MoO,CI,. WO,CI,, and Ce(OH)(N0,),.3H,O (ICN powder camel2 (diameter 57.3 mm). Each sample was analyzed Canatla Lttl.). ErCI,.HZO 21nd YCI,.H,O were made by soluhilizing El.,(), and YLOj (Koch-Light l.2iibo1.i~tory.Ltd.. using suitable filtered radiation (e.g. F e K alpha radiation for Colnbrook. Bucks, Engl21ntl)respectively in dilute HCI (ratio of gold) and identification obtained by comparison with the Joint Committee on Powder Diffraction Standards file (1974) for each niet:~l:chlo~-ide was I:?) and heating to dryness. All salts were ailded to deionized distilled water (DDW) to make 10 m ~ \ d particular element. solutions of each metal immediately before the experiment. Elcc./rorr itlic.rosc.o~/~.~ oJ' Wolls WIric.11 W c ~ , rRetrc~teclii.irlr tlrc, Solutions which were light-sensitive (i.e. AuCI,.2H,O, AgNO,. Vtrr.iorr.s Srtritrirr,y Kecr,yrrrts UO,(C,H,O,),.2H,O. etc.) were kept in the dark 21s were the One millilitre of a cell wall suspension (1.0 mg dry wt. per experiments with them performed in the dark. Some solutions millilitre) was reacted with 1 ml of the various concentl.ations of (e.g. HgCI,) required heat to facilitate easy solubility whereas the stains for varying periods of time (see Table 3). For the others (e.g. Bi(NO;),.5H,O) requil-ed the ;iddition of acid. 0s-meth reaction a 1.0-mg quantity of freeze-dried cell walls Elcc.irotr-~tric~rosc~o/)ic Sitritrs was added to 2 ml of the 0.25% solution of the staining reagent. The following compounds were used as electron-cnicroscopic These specimens were then processed as describedfor the metal stains: 2% uranyl acetate (Fisher Scientific Co.), 2% osmium salt reactions. Quantitative estimations of the amounts of the g red per millilitre. tetroxide (BDH Chemicals), 5 0 0 ~ 1ruthenium stains absorbed by the walls were established by analysis for the 0.25% 0s-meth (an osmium tetroxide - hexarnethylene- heavy atomic number metal contained by each staining reagent tetramine complex). a 10 mM and a saturated (about 0.1%) (e.g. detection of the Ruin ruthenium red). aqueous solution of 0s-DMEDA (an osmium-dimethylethylenediamine complex). 2% uranyl magnesium acetate (all Frrc.:,c-ctc.lrirr!: c~f'tlreCell Wtrlls from Polysciences Inc.), and 10 mM KPtCI,. DMSO (a complex One milligram dry wt. of wall material was infiltrated with an of potassium, platinum, chlorine, and dimethylsulfoxide. which aqueous 10% solution of glycerol. This materi~ilwas freezewas kindly supplied by T. Jack. Department of Chemistry, cleaved, etched for 30-60 s, platinum-shadowed, and carbonScarborough College, University of Toronto. Canada). replicated in a Balzers model BA5lOM freeze-etching apThe 0s-meth solution was prepared by dissolving 25 mg of the paratus. compound in dimethylformamide and then adding a sufficient Ulirerc~qocorrt~ of rlrt, Cell Wtrlls volume of DDW to bring the total volume to 10 ml. One-milligram (dry wt.) samples of the cell walls were susIn s e v e ~ ainstances.after l the cell walls had been subjected t o a pended in 1% 'curanyl acetate for 10 rnin, either infiltrated with metal salt solutionof low electron-scattering powel'. asolutionof 20% bovine serum albumin (BSA - Nutritional Biochemicals ferritin conjugated to concanavalin A (F-ConA) was applied to Corp.) or 1.1 M sucrose (ribonuclease free-SchwartzlMann, the walls to act as a wall-surfrice marker ( I ) . Division of Becton. Dickinson and Co.), frozen in liquid nitroElec~/r.or~ Mic,~,o.sc,opy qf'Cc,ll Wcrlls Wlricl~Werr Srrl?jt~c~lc~el lo t l ~ c gen, and sectioned with a Reichert model OMU3 ultramicroM~rcrlStrll So1rrtiotr.s tome equipped with a model FC2 ultracryotomy unit. T h e BSA One millilitre of a cell wall suspension (1.0 mg dry wt. per sample was cross-linked with 1% glutaraldehyde for 10 min a t millilitre) was mixed I:]with each 10 mM metal salt solution to 22°C and well-washed before freezing. Both the BSA and the give a final metal concentration in the reaction mixture of 5 mM sugar were used a s agents to control the hardness of the frozen and the reaction was allowed to proceed for 10 min at 22°C. The samples (18). walls were pelleted by centrifugation and washed with 5 volumes of 10 ml DDW after which they were fixed in 4% aqueous Dit~r~rrsiotrol Atro1ysc.s c~f'7'lrirrSectiotrs c(f'tlrt, Ccll WtrIls glutaraldehyde (Polysciences, Inc.) for 30 min at 22'C. The At least 30 sepal-ate and random measurements were taken walls were then well-washed, enrobed in agar, dehydrated from micrographs (magnification between 200000 and 300000 through an eth;inol+propylene oxide series, and embedded in X ) of cross sections (both transverse and longitudinal cuts) of Epon 812. Cell walls which were reacted with low atomic the walls which had been treated with each of the metal salts o r number (2)elements were often treated with F-ConA imstaining reagents. From this data :i 'mean' thickness (p). a mediately before fixing and embedding ( I ) . standard deviation ( a , ) and , a standa1.d error (SE) was calculated. All blocks were sectioned with a Reichert model OMU2 T h e dimensional aspects of the walls, which were reacted with ultr;~microtomeand the sections (about 60 nm in thickness) metals that provided a high degree of electron scattering, were collected on carbon-Formvar-coated 200-mesh grids. No stains evaluated using Student's 1 distribution and significant ( ~ / ( ~ . 7 0 Q r r t r r r ~ i ~ t r ~ i ~ ~ t ~ A t r t(