Dear Pengfei Ding,

I have used a special EM  stain,long ago,to stain nucleic acids differentially with regard to their DNA/RNA composition by Bismuth staining. It is not widely known, but I will try to provide you with some links. Kind regards, Harrie Verhoeven.

Uranyl Acetate (UA) is the best staining material for nucleic acids. The advantage of UA is that produces the highest electron density and image contrast as well as imparting a fine grain to the image due to the atomic weight of 238 of uranium. The uranyl ions bind to proteins and lipids with sialic acid carboxyl groups such as glycoproteins and ganglioside and to nucleic acid phosphate groups of DNA and RNA.

The attachment will guide you to prepare a staining solution.

While my cherished colleagues above posted good solutions (and naturally UO2-acetate is the most important candidate for nucleic acids (mainly for DNA)  in double staining, also in its other chemical forms often applied to ultrathin sections, namely uranyl-formate or uranyl magnesium acetate (7,5 % for instance), I would like to add at least also the link to another RG-Question asked & answered multiplicatively 3 years ago: cf.: 

https://www.researchgate.net/post/What_is_the_last_method_for_staining_of_DNA_and_DNA_detection_in_ultra_structure_samples_for_Electron_Microscopy_EM .

Michael J. Dykstra, ‎Laura E. Reuss - 2003 in their book: Biological Electron Microscopy: Theory, Techniques, and Troubleshooting  describe  ".....However, RNA is more reactive with lead stains"......                                                            @cf: https://books.google.at/books?isbn=0306477491".

Further candidates:

(phosphotungstic acid, phosphomolybdic acid), "EDTA-regressive" staining for ribonucleoproteins results in chromatin bleaching and stains RNA-containing Structures (Bernhard W,1969, A new staining procedure for electron microscopical cytology. J. Ultrastruct.Res.27,250-265, ARTICLE available only via PPV=Pay Per View:  cf:  http://www.sciencedirect.com/science/article/pii/S002253206980016X ).

Ethidium bromide- and propidium iodide-PTA staining for the demonstration of RNA-containing structures (Biggiogera and Flach-Biggiogera 1989, Ethidium bromide- and propidium iodide-PTA staining of nucleic acids at the electron microscopy level. J.Histochem.Cytochem 37, 1161-1166. Abstract: http://journals.sagepub.com/doi/abs/10.1177/37.7.2471727, PDF @: http://journals.sagepub.com/doi/pdf/10.1177/37.7.2471727 

Bromination of (LR-White) ultrathin sections (mounted on nickel grids) followed by specific antibodies against bromo-deoxyuridine might be another way: cf: 

CALABUIG et al, Specific detection of RNA on ultra-thin sections, Journal of Structural Biology 152 (2005) 146–148:

Abstract (only for convenience)

In this paper, we describe a specific method for ultrastructural detection of RNA. Our method is based in the bromination ‘‘in situ’’ of uridine residues in the RNA, which allows the detection of brominated RNA with specific antibodies against bromo-deoxyuridine. With this method we can achieve high specificity and resolution, and it can be applied to Epon or acrylic resin embedded material.

(C) 2005 Published by Elsevier Inc. Article Abstract cf.:  http://www.sciencedirect.com/science/article/pii/S1047847705001425 ; PDF available only via PPV.

Keywords: Electron microscopy; Nucleic acid detection; RNA

EM-Staining for DNA and RNA have been reviewed (only one example out of some other references) by LEWIS PR and KNIGHT DP 1977: Staining Methods for sectioned material. In series: GLAUERT A.M.(Ed), Practical Methods in Electron Microscopy, Vol.5, Part I, pp. 1-311, Elsevier / North Holland Biomedical Press. Amsterdam (NL).

It might be worth thinking about the influence of (different) fixatives usually applied to tissue(s) for analytical TEM(positive stain/contrast )

Best wishes and good luck!

NB: This answer was edited for /with the links to mentioned articles in respective journals 50 min afetr initial posting.

Dear Pengfei Ding:   compared to my Re# 03 above:

'Gold rush' perhaps the favour of the day.... (:-)) !

Interestingly, the most promising technique for (a specific) RNA-stain in / for (T)EM I was able to find in my e-files / e-library right now*): 

BIGGIOGERA M. and FAKAN St.,1998:

"Fine Structural Specific Visualization of RNA on Ultrathin Sections" in Journal of Histochemistry and Cytochemistry Vol. 46(3): 389–395.

In this article, the preparation of the reagent, terbium citrateß), the staining with the special reagent 'C6H5O7Tb' on expoxy/epon as well as acrylic sections is described.

Sincerest THANKS to the SAGE§) Organisation (https://us.sagepub.com/)  for offering free open access to most of the OLD editions/volumes/issues/articles of e. g. JHC!  

SUMMARY (only for convenience of the reader): "We describe a new technique that allows specific visualization of RNA at the electron microscopic level by means of terbium citrate. Under the conditions presented here, terbium binds selectively to RNA and stains nucleoli, interchromatin granules, perichromatin fibrils, perichromatin granules, and coiled bodies in the cell nucleus, whereas ribosomes are the only contrasted structures in the cytoplasm. All the cell components contrasted by terbium are known to contain RNA. When ultrathin sections are [note added: CONTROLS]  pretreated with RNase A or nuclease S1 (specific for single-stranded nucleic acids), staining does not occur.

Neither DNase nor pronase influences the reaction. We conclude that terbium staining is selective for RNA and especially for single-stranded RNA. The staining can be performed on thin sections of material embedded both in epoxy and in acrylic resins. The technique is not influenced by the aldehyde fixative used and can also be utilized after immunolabeling [note added: in this case with a monoclonal anti-DNA antibody]. The endproduct is very fine and, although weak in contrast, is suitable for high-resolution observations."

In addition to the convenience of including the SUMMARY:

This 'old but gold' excellent technical article can be found  @: journals.sagepub.com/doi/pdf/10.1177/002215549804600313 

More on Terbium citrate cf. e.g.: https://pubchem.ncbi.nlm.nih.gov/compound/202905

C6H5O7Tb: Molecular Weight: 351.06 - Boiling Point: 309.6 °C at 760 mmHg  -

Flash Point: 155.2 °C 

footnotes:

*) due to loss of my former, almost ready for posting new answer text it took another 20 min to edit and complete this swecond version of my answer. Hope you are satisfied with....

 ß) CAVE/CAUTION/SAFETY indications:  Hazard code: Xi: Terbium citrate  is a poison by intraperitoneal route. When heated to decomposition it emits acrid smoke and irritating fumes. Water pollutant class III (in Germany: WGK3). Special safety measures (in lab and handling may apply  - inter alia: 'Wear suitable protective clothing' ).

§) honestly to be mentioned:

Sara Miller McCune founded SAGE Publishing in 1965 to support the dissemination of usable knowledge and educate a global community.

SAGE Publishing is an independent publishing company and today based in California, USA.

Disclaimer: no affiliation to and no financial interest in any mentioned company or potential dealer of the mentioned  reagent(s).