Journal of Electron Microscopy 59(2): 113118 (2010) doi: 10.

1093/jmicro/dfp045
.........................................................................................................................................................................................................................................................

Biological: Full-length

Examination of electron stains as a substitute for uranyl acetate for the ultrathin sections of bacterial cells
Kaoru Yamaguchi , Ken-ichiro Suzuki and Kenji Tanaka
Biological Resource Center (NBRC), Department of Biotechnology, National Institute of Technology and Evaluation, 2-5-8 Kazusakamatari, Kisarazu, Chiba 292-0818, Japan To whom correspondence should be addressed. E-mail: yamaguchi-kaoru@nite.go.jp
................................................................................................................................................................................................

Abstract

Electron staining reagents were examined to nd a possible substitute for uranyl acetate (UA) in electron microscopy of bacterial ultrathin sections. Four kinds of stains, platinum blue (Pt-blue), oolong tea extract (OTE), potassium permanganate (KMnO4 ) and phosphotungstic acid (PTA), were examined in comparison with UA either with or without post-staining with lead citrate (Pb). Electron microscopy was performed on sections from Spurr-embedded cells of a Gram-positive bacterium, Bacillus cereus NBRC 13597, and a Gram-negative bacterium, Escherichia coli NBRC 3301. Both Pt-blue and OTE showed staining similar to each other and to that of double staining with UA and Pb in B. cereus, while in E. coli the cytoplasmic membrane appeared less dense when compared with UA and Pb. KMnO4 stained excessively to some extent, but showed images of the best contrast in the cytoplasmic membrane comparable with UA and Pb among the four reagents. PTA could stain the peptidoglycan layer but gave images of low quality for both bacteria. This study demonstrated that none of the reagents examined showed staining results of the same quality or better than the conventional method with UA and Pb. However, stains of Pt-blue, OTE and KMnO4 could possibly be an alternative candidate for the UA according to the structure in question. electron staining, bacteria, freeze substitution, platinum blue, oolong tea extract, potassium permanganate 2 March 2009, accepted 13 August 2009, online 19 September 2009

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

................................................................................................................................................................................................

Keywords Received

................................................................................................................................................................................................ ................................................................................................................................................................................................

Introduction
For the ultrastructural study by transmission electron microscopy (TEM), ultrathin sections have been stained with an aqueous solution of uranyl acetate (UA) followed by lead stains. This protocol of double staining has been conventionally used as the most useful and promising staining method. However, the recent trend that buying and use of uranyl compounds are strictly regulated by the government as they are internationally controlled nuclear materials has made it difcult to use UA as an electron stain for conventional electron microscopy. Under these circumstances, it is quite reasonable to re-examine various stains once attempted years ago [13] and try the

newly reported compounds, platinum blue (Pt-blue) [4] and oolong tea extract (OTE) [5] as a substitute of UA for electron staining. In this study, we examined various stains on the ultrathin sections of bacterial cells xed by the freeze-substitution method. Staining effects of the different stains on the sections stained either singly or with the following lead citrate (Pb) staining were compared.

Materials and methods

A Gram-positive bacterium, Bacillus cereus NBRC 13597, and a Gram-negative bacterium, Escherichia coli NBRC 3301, were used. Bacterial cells were cultured at 30 C on an agar plate containing

.........................................................................................................................................................................................................................................................

c The Author 2009. Published by Oxford University Press [on behalf of Japanese Society of Microscopy]. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

114

J O U R N A L O F E L E C T R O N M I C R O S C O P Y , Vol. 59, No. 2, 2010

polypeptone 1% (Wako Pure Chemical Industries, Osaka, Japan), yeast extract 0.2%, MgSO4 7H2 O 0.1% and agar 1.5% at pH 7.0. Cells cultured overnight were inoculated into the fresh agar medium and cultured for 34 h at 30 C. Growing cells were picked up from the plate and sandwiched between two copper grids, which were xed by rapidly plunging into liquid propane cooled in liquid nitrogen using Leica EM CPC (Leica Microsystems, Wetzlar, Hessen, Germany) and transferred to a 2% osmium tetroxide solution in absolute acetone at 80 C in Leica EM AFS (Leica Microsystems). Samples were embedded in Spurrs resin [6]. Sections were obtained with a diamond knife on the ultramicrotome Ultracut UCT (Leica Microsystems), and picked up on the formvarcoated grids. Sections on grids were stained by oating them on single drops of the staining solution at room temperature following one of the protocols (1) to (5) described below, rinsed in deionized water and dried and then, a part of the sections were directly subject to observation (single staining), while the others were post-stained with the Pb solution for 5 min (double staining): (1) 3% aqueous solution of UA for 2 h. (2) Pt-blue solution (supplied by courtesy of Dr Inaga, Tottori University, and now it is possible to purchase it from Nisshin-EM, Tokyo, Japan) diluted 10 times with deionized water for 10 min [4]. (3) 0.2% OTE (Nisshin-EM) in a 1/15 M potassium phosphate buffer, pH 7.4, for 60 min [5,7]. (4) 0.5% potassium permanganate (KMnO4 ) in a 1/15 M potassium phosphate buffer, pH 7.0, for 1 min, followed by a rinse with a 0.05% citric acid solution for 10 s [1,2]. Note that a resin including methyl nadic anhydride (MNA) is not recommended, as it is darkened by KMnO4 [8]. (5) 5% aqueous solution of phosphotungstic acid (PTA) for 60 min [3]. (6) None, or only Pb staining for 5 min [9]. Sections were observed by Hitachi H-7600 electron microscope at 100 kV. The images were taken using the AMT CCD digital camera system (Advanced Microscopy Techniques, Danvers, MA, USA).

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

Fig. 1. Ultrathin sections of Bacillus cereus (a) and Escherichia coli (b) xed by the freeze-substitution method stained with UA + Pb. The ribosomes and nuclear materials were uniformly distributed in the cytoplasm. Bar (a and b) 500 nm.

Results
The ultrastructure of the bacterial cells prepared by the freeze-substitution method and stained using UA and Pb is shown in Figs. 1a, b, 2a and 3a. It is characterized by uniformly distributed ribosomes and nuclear materials in the cytoplasm (Fig. 1a and b), in contrast with the images obtained by conventional aldehyde-osmium xation where the nuclear region is found in the middle less dense area demarcated from diffused ribosomes in the cytoplasm [10,11]. The cell wall and the cytoplasmic membrane are tightly appressed in freeze-substituted cells (Figs. 2a and 3a), while both structures are loosely separated in conventionally xed cells [10,11]. In this study, we described mainly the results about the contrast of the cell wall and the cytoplasmic membrane, not the cytoplasmic structures which were not so clearly apparent as to distinguish the details in contrast between the different staining methods. The appearance of the surface structure is markedly different between Gram-positive and Gram-negative bacteria [10,11]. This difference could be seen in the sections by digital image acquisition without post-staining (single staining), but markedly in the sections treated with post-staining by Pb (double staining). The cell wall of Gram-positive B. cereus is composed of a thick peptidoglycan layer (Fig. 2a) and the surface structure of Gram-negative E. coli is

K. Yamaguchi et al. Examination of electron stains for the bacteria

115

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

Fig. 2. Ultrathin sections of Bacillus cereus xed by the freezesubstitution method stained with (a) UA + Pb, (a ) UA, (b) Pt-blue + Pb, (b ) Pt-blue, (c) OTE + Pb, (c ) OTE, (d) KMnO4 + Pb, (d ) KMnO4 , (e) PTA + Pb, (e ) PTA, (f ) Pb and (f ) no stain. Cell wall including peptidoglycan (CW) and cytoplasmic membrane (CM) were stained clearly to the similar degree of contrast with all but PTA reagents by post-staining with Pb. With PTA, CW was densely stained. Bar (af ) 100 nm.

Fig. 3. Ultrathin sections of Escherichia coli xed by the freezesubstitution method stained with (a) UA + Pb, (a ) UA, (b) Pt-blue + Pb, (b ) Pt-blue, (c) OTE + Pb, (c ) OTE, (d) KMnO4 + Pb, (d ) KMnO4 , (e) PTA + Pb (e ) PTA, (f ) Pb and (f ) no stain. Outer membrane (OM) and peptidoglycan (PG) were distinctly stained with all reagents by post-staining with Pb. On the other hand, cytoplasmic membrane (CM) was distinctly demonstrated with only UA and KMnO4 . Bar (af ) 100 nm.

116

J O U R N A L O F E L E C T R O N M I C R O S C O P Y , Vol. 59, No. 2, 2010

multilayered, with a layer of peptidoglycan between the outer membrane and inner cytoplasmic membrane (Fig. 3a). The results obtained with the different staining methods are described here in comparison with those using double staining with UA and Pb (Figs. 2a and 3a). In B. cereus, Pt-blue stained clearly both the cell wall and the cytoplasmic membrane (Fig. 2b). The cell wall of B. cereus was stained variably with the different stains: Pt-blue and OTE stained similarly (Fig. 2b and c), and PTA stained densely (Fig. 2e). In E. coli, Pt-blue stained the outer membrane well, but less dense the peptidoglycan layer and the inner cytoplsmic membrane (Fig. 3b). Similar images were obtained in both kinds of OTE and PTA staining (Fig. 3c and e). Pt-blue and OTE by single staining hardly exhibited contrast in both cells (Figs. 2b , c and 3b , c ). On the other hand, single staining with PTA provided a contrast in structure, especially the peptidoglycan layer of the cell wall (Figs. 2e and 3e ). A diluted solution of KMnO4 stained more intensely the ultrastructure of the cell wall revealing the details to the same extent as that obtained with UA in both bacteria, but produced rough images of the cytoplasm suggesting a brief damage (Figs. 2d and 3d). KMnO4 alone could provide images similar to those obtained on single staining with UA but with a reduced contrast (Figs. 2d and 3d ). Although sections either stained singly with Pb or with no staining were hardly observed on the uorescent screen under the electron microscope, the images were acquired through CCD followed by image processing with the AMT system. Therefore, the ultrastructural details on sections stained singly with Pb could be seen in some contrast comparable to those obtained by other double staining protocols (Figs. 2f and 3f ), but no structure could be discerned for those with no staining treatment (Figs. 2f and 3f ). Above results are summarized in Table 1.

Discussion
We examined various kinds of electron stains as a possible substitute for UA in the double staining method with Pb for sections of bacterial cells. In addition, to compare the staining effects of the differ-

ent stains with that of UA, the images were obtained on the sections treated either with the staining solution alone (single staining) or with that combined with the following Pb-staining (double staining). Our studies showed that most of the stains used could reveal overall structures, but were different in some details. And we concluded that double staining with UA and Pb used generally in conventional TEM observations gave the best contrast of the structural details. The uptake of the stain depends on many factors which include xative, embedding medium, the nature of the stain and the vehicle in which it is dissolved, and also the concentration, temperature and pH of the staining solution. In this study, we followed the staining conditions described in the references [4,9] and in some cases modied so as to obtain good staining results [13,5,7]. It was shown that both Pt-blue andOTE could be useful as an alternative stain to UA, although the obtained images were partly in the less contrast. Pt-blue that was prepared from reactions of cisdichlorodiamine-platinum (II) (cisplatin) with silver nitrate and then with thymidine has been introduced for the positive staining of ultrathin sections and provided images of virus and animal cells in good contrast when followed with post-staining with Pb by Inaga et al. [4]. In consistent with their observations, our results also showed the possibility to be applicable to the TEM of bacterial cells. The enhanced contrast with Pt-blue was produced by post-staining with Pb, as TEM after single staining showed no contrast. Sato et al. [7] introduced OTE for TEM of connective tissues, and found that it enhanced the contrast of elastin and collagen brils. And they recommended OTE generally as a substitute for UA in staining ultrathin sections [5]. OTE contains polyphenols similar to tannic acid, which is employed as an adjunct to aldehyde xation. The use of tannic acid after routine aldehyde-osmium xation, or its inclusion in the xative can enhance the contrast. Also, treatment of the sections with aqueous tannic acid before staining has often been performed [12]. From these observations for tannic acid treatment, it can be concluded that it is reasonable to use OTE for the enhancement of contrast by staining, and our results seem to be consistent with

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

K. Yamaguchi et al. Examination of electron stains for the bacteria

117

.........................................................................................................................................................................................................................................................

Table 1. Comparison of various stains for bacterial cells Bacillus cereus Electron staining reagent UA + Pb Pt-blue + Pb OTE + Pb KMnO4 + Pb PTA + Pb Pb Cell wall ++ ++ ++ ++ +++ ++ Cytoplasmic membrane ++ + + ++ + + Escherichia coli Outer membrane ++ ++ ++ ++ ++ ++ Cytoplasmic membrane ++ ++

Cytoplasm ++ ++ ++ +++ ++ ++

Figure 1a, 2a 2b 2c 2d 2e 2f

Peptidoglycan ++ + + ++ + +

Cytoplasm ++ ++ ++ +++ ++ ++

Figure 1b, 3a 3b 3c 3d 3e 3f

, not stained; +, weakly stained; + +, moderately stained; + + +, strongly stained.

those obtained for tannic acid treatment, that is, cell wall structures were clearly demonstrated after the post-staining method, while no structure could be discerned by single staining treatment with OTE. KMnO4 had once been introduced as a xative by Luft [13], and then had been used for plant cells by Mollenhauer [14]. KMnO4 yielded the membranous structures in good contrast, but failed to reveal the ne details such as ribosomes of the cytoplasm [13,14]. Yasutake and Murakami [2] used KMnO4 as a staining reagent for the sections of animal cells embedded in methacrylate resin and demonstrated the high-contrast images when followed with poststaining with Pb. Similar observations using Spurrs resin [6] were made by Bray and Wagenaar [1], which also emphasized the effects of post-staining with the Pb solution. We could also show that a diluted solution of KMnO4 alone could give a little contrast, but double staining with Pb gave the best contrast for the bacterial cell wall examined in this study. PTA was widely used as a stain for negative staining, but it was rarely used for positive staining, for either of the sections with 2% PTA made up in 90% ethanol [15], or en bloc staining before embedding after aldehyde xation [16]. Watson [3] applied a 10% solution of PTA on the methacrylate sections and found it effective to stain collagen heavily. Furthermore, our results using an aqueous solution of 5% PTA alone also demonstrated that PTA was effective for increasing the contrast of the peptidoglycan layer of the cell wall structure, but the images obtained by the double staining showed low quality for both bacteria.

terial cells when combined with post-staining with Pb. And even images with a very low contrast on the uorescent screen in TEM, such as the images obtained either by no staining or by single staining with Pb, could be enhanced by digital processing of the contrast. The images obtained by conventional staining with a combination of UA and Pb revealed distinctly the structural details of the cell wall and cytoplasm. Generally, Pt-blue, OTE and PTA gave the images of a lower quality to show the ultrastructure of the cell surface. However, KMnO4 could show the structural details almost to the same level as UA, and might acquire better images if the staining conditions could be improved. Observations on the sections with single staining could reveal the difference in the selective staining properties among the stains used here. KMnO4 and PTA stained some structures which are enhancingly demonstrated by the second Pb staining, while Pt-blue and OTE seemed to act as a mordant giving no structures by single staining which are revealed after post-staining with Pb for bacterial cells.

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

Acknowledgements
We sincerely thank Dr S. Inaga, Division of Genome Morphology, Department of Functional, Morphological and Regulatory Science, Faculty of Medicine, Tottori University, for providing an aliquot of the Pt-blue solution and advice on the staining method using Pt-blue. Also thanks to Dr A. Hirata, Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, for performing the staining with the UA solution. We thank Dr K. Mori, NBRC, and Dr T. Iino, Japan Collection of Microorganisms (JCM), RIKEN Bioresource Center, for providing advice on this study.

Concluding remarks
All kinds of stains we examined could give an almost satisfactory contrast for the sections of bac-

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...............................
1 Bray D F and Wagenaar E B (1978) A double staining technique for improved contrast of thin sections from Spurr-embedded tissue. Can. J. Bot. 56: 129132.

118

J O U R N A L O F E L E C T R O N M I C R O S C O P Y , Vol. 59, No. 2, 2010

........................................................................................................................

........................................................................................................................

2 Yasutake S, Murakami M, and Yoshida N (1961) Double staining of tissue sections for electron microscopy with heavy metals (KMnO4 and Pb(CH3 COO)2 ). J. Electron Microsc. 10: 233237. 3 Watson M L (1958) Staining of tissue sections for electron microscopy with heavy metals. J. Biophys. Biochem. Cytol. 4: 475478. 4 Inaga S, Katsumoto T, Tanaka K, Kameie T, Nakane H, and Naguro T (2007) Platinum blue as an alternative to uranyl acetate for staining in transmission electron microscopy. Arch. Histol. Cytol. 70: 4349. 5 Sato S, Adachi A, Sasaki Y, and Ghazizadeh M (2007) Oolong tea extract as a substitute for uranyl acetate in staining of ultrathin sections. J. Microsc. 229: 1720. 6 Spurr A R (1969) A low viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res. 26: 3143. 7 Sato S, Sasaki Y, Adachi A, Dai W, and Liu X-L (2003) Use of oolong tea extract (OTE) for elastin staining and enhancement in ultrathin sections. Med. Electron Microsc. 36: 179182. 8 Reedy M K (1965) Section staining for electron microscopy. Incompatibility of methyl nadic anhydride with permanganates. J. Cell Biol. 26: 309311. 9 Reynold E S (1963) The use of lead citrate at high pH as an electronopaque strain in electron microscopy. J. Cell Biol. 17: 208212.

........................................................................................................................

........................................................................................................................

........................................................................................................................

........................................................................................................................

........................................................................................................................

........................................................................................................................

Downloaded from http://jmicro.oxfordjournals.org/ by guest on May 13, 2012

........................................................................................................................

10 Amako K, Murata K, and Umeda A (1983) Structure of the envelope of Escherichia coli observed by the rapid-freezing and substitution xation method. Microbiol. Immunol. 27: 9599.
........................................................................................................................

11 Amako K and Takade A (1985) The ne structure of Bacillus subtilis revealed by the rapid-freezing and substitution-xation method. J. Electron Microsc. 34: 1317.
........................................................................................................................

12 Knight D P and Lewis P R (1992) Cytological staining methods, other sytological stains, other non-specic staining methods. In: Lewis P R and Knight D P (eds), Cytochemical Staining Methods for Electron Microscopy, pp 64, 65 (Elsevier, Amsterdam).
........................................................................................................................

13 Luft J H (1956) Permanganate; a new xative for electron microscopy. J. Biophys. Biochem. Cytol. 2: 799802.
........................................................................................................................

14 Mollenhauer H H (1959) Permanganate xation of plant cells. J. Biophys. Biochem. Cytol. 6: 431436.
........................................................................................................................

15 Richardson W D and Davies H G (1980) Quantitative observations on the kinetics and mechanisims of binding of electron stains to thin sections through hen erythrocytes. J. Cell Sci. 46: 253278.
........................................................................................................................

16 Silverman L and Glick D (1969) The reactivity and staining of tissue proteins with phosphotungstic acid. J. Cell Boil. 40: 761767.