Methodology of Protistan Discovery: from rRNA Detection to Quality Scanning Electron Microscope Images{dagger}

Each year, thousands of new protistan 18S rRNA sequences aredetected in environmental samples. Many of these sequences aremolecular signatures of new protistan species, classes, and/orkingdoms that have never been seen before. The main goal ofthis study was to enable visualization of these novel organismsand to conduct quality ultrastructural examination. We achievedthis goal by modifying standard procedures for cell fixation,fluorescence in situ hybridization, and scanning electron microscopy(SEM) and by making these methodologies work in concert. Asa result, the same individual cell can now be detected by 18SrRNA-targeted fluorochrome-labeled probes and then viewed bySEM to reveal its diagnostic morphological characteristics.The method was successfully tested on a wide range of protists(alveolates, stramenopiles, kinetoplastids, and cryptomonads).The new methodology thus opens a way for fine microscopy studiesof many organisms previously known exclusively by their 18SrRNA sequences.

INTRODUCTION

Top

Introduction

Research on protistan diversity started two centuries ago (5)but still remains in its infancy. The application of moleculartools to environmental samples is producing a seemingly endlesslist of organisms that are apparently missed by classical approaches(3, 4, 12, 14, 15, 20). The high rate of acquiring novel protistanrRNA sequences suggests that the global protistan inventoryis utterly incomplete and that the captured diversity is butthe

tip of an iceberg.

The novelty of some of these sequencesindicates discovery of protistan lineages at all levels of taxonomichierarchy, including new kingdoms (3). There is little doubtthat entirely new classes of Protista remain undetected, unseen,and unexplored.

The rRNA approach (2) has proved to be uniquely suited for theinitial discovery of novel organisms but provides little informationbeyond the fact of their existence, abundance in nature, andmolecular phylogeny (13). The organisms detected remain bewildering,and their basic biology stays secret. Only direct access tothe novel organisms would validate their assignment to hightaxonomic levels (3), arguments about the ecological significanceof their diversity (13), and their use in reconstruction ofearly eukaryotic evolution (17, 18). Gaining access to the novelorganisms is therefore a natural next step after the discoveryof their molecular signatures.

Unlike prokaryotes, microscopic eukaryotes exhibit a very highlevel of morphological diversity. Because of this, light andelectron microscopy studies of the newly discovered organismswould be helpful for confirming their uniqueness, assessingtheir relatedness to other known protists, and unraveling certainaspects of their lifestyles. However, currently available methodsdo not allow such studies.

This paper describes a new combination of the modified 18S rRNA,fluorescence in situ hybridization (FISH), and scanning electronmicroscopy (SEM) methodologies. This combination allows identificationof target cells by FISH to be followed by fine-quality SEM studies,with both performed on the same microscopic preparations. Thisopens a way for researchers to conduct proper morphologicalstudies of organisms known today only from their 18S rRNA signaturesin the environment.

MATERIALS AND METHODS

Top

Materials and Methods

Test organisms.
The test species used in this study included representativesof several major protistan taxa, such as alveolates, stramenopiles,kinetoplastids, and cryptomonads (Table 1). Cultures of allspecies were obtained from Carolina Biological Supply (Burlington,N.C.), except the ciliate Euplotes sp., which was isolated fromthe Sippewisset salt marsh, Cape Cod, Mass.

View this table:
[in this window]
[in a new window]
TABLE 1. Classification of strains used for this study

Fixation.
We tested several standard and modified fixative mixtures (11)for the ability to impart sufficient rigidity to the test cells.The list included paraformaldehyde (tested at 2, 4, and 10%;also tested at 2% supplemented with 0.2% glutaraldehyde), formalin(tested at the same concentrations and combinations as describedfor paraformaldehyde), Schaudinn's fixative (2 parts saturatedaqueous HgCl₂, 1 part 100% ethanol, and 1% [final concentration]glacial acetic acid; 1 part of this fixative was diluted with3 parts distilled water and added to the sample at a ratio of1:1 [vol/vol]), Steve's fixative (76% saturated aqueous HgCl₂,20% formaldehyde, and 4% glacial acetic acid mixed with 3 partsdistilled water and added to the sample at a ratio of 1:1 [vol/vol]),and Bouin's fixative mixture (buffered formaldehyde saturatedwith picric acid and 2% glacial acetic acid added immediatelybefore fixation) tested at 50% with or without the additionof 0.1 to 0.2% glutaraldehyde. The fixation time in all caseswas 45 min at 4°C.

FISH probes and probe design.
FISH was used to visualize the marine ciliate Euplotes sp. bothin environmental samples and mixed with other test protists(Table 1) as negative controls. We designed two oligonucleotideprobes to target different regions of the 18S rRNA. A set ofprobes (nucleotides, 18 to 22; GC contents, 50 to 60%; nucleotide-nucleotideT_ms, $>=$ 57°C) was designed for each Euplotessmall-subunit (SSU) rRNA sequence retrieved from GenBank byuse of the online tool Primer3 (19). Generated probes were checkedagainst the GenBank sequence collection by a standard nucleotide-nucleotideBLAST search (1) and were compared to an accessibility map ofthe SSU rRNA of Escherichia coli for hints of probe target siteswith promising high signal intensities (8). The potential forhairpin and dimer formation of selected oligonucleotides wasassessed by use of the program mfold, v. 3.0 (21). From theoriginal 54 probe candidates, two oligonucleotides were chosenthat fulfilled the general criteria of potentially successfulprobes (10, 16). The probe Eupl240 (5'-TCATCTCAGTAGACCTTGCG-3')had no mismatches with the SSU rRNA of all but one Euplotesspecies (E. raikovi) but exhibited a 3-bp mismatch with thenext closest sequence in GenBank. The probe Eupl1780 (5'-GACAGTCCAAAGAGGTTCAC-3')matched all 33 Euplotes sp. SSU rRNA sequences deposited inGenBank and had a mismatch of 3 bp with the next closest GenBanksequence. The probes used were purified by high-performanceliquid chromatography and labeled with Cy-3 at the 5' end bythe manufacturer (Integrated DNA Technologies, Coralville, Iowa).

Other probes used included the universal Cy-3-labeled eukaryoticprobe Euk1209R (5'-GGGCATCACAGACCTG-3') (9) and its Cy-3-labeledcomplement as a nonsense probe.

FISH staining.
We used a standard protocol for in situ hybridization (16),with several important modifications. Preserved samples wereplaced in wells with bottoms made of polycarbonate membranes(25-mm diameter; 0.4-µm pore size) (Costar Transwells;Corning Inc., Corning, N.Y.). The wells were placed on the baseof standard 25-mm-diameter glass filtration units (Millipore,Bedford, Mass.) (Fig. 1). To ensure an even cell distributionon the well's membrane, a 5.0-µm-pore-size 25-mm-diametercellulose nitrate filter (Whatman, Newton, Mass.) was insertedbetween the well's bottom and the glass base. The excess offixative was removed by gentle filtration (<200 mm Hg), alwaysleaving a small amount of liquid covering the membrane. Thecells were washed by gradually replacing the remains of thefixative with 1x phosphate-buffered saline (PBS) (0.14 M NaCl,2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.4 mM KH₂PO4, pH 7.4). This wasachieved in three to five cycles (in the case of Bouin's fixative,washing was continued until the yellow coloration disappeared),each consisting of adding 2 ml of fresh PBS and gently removingit by applying a weak vacuum. We found it essential for thepreservation of cell shape to never expose the membrane surface(and thus cells) to drying. The washing buffer was then graduallyreplaced with a hybridization buffer (0.9 M NaCl, 20 mM Tris-HCl[pH 8], 30% formamide [the optimal concentration was found inpreliminary experiments], and 0.01% sodium dodecyl sulfate).This was done in two cycles, each consisting of adding and thenremoving 2 ml of the fresh buffer. The wells, with approximately300 µl of hybridization buffer, were transferred to asix-well Transwell tissue culture tray, and 50 µl of probesolution (30 ng µl^-1 in molecular-grade distilled water)was added to each well. The wells received either a universaleukaryotic, nonsense, or Euplotes-specific probe (separately).The wells with probes added were transferred into glass inserts(stacking Stender dish; 27-mm inner diameter; volume, 8 ml),which were placed in a tightly sealed incubation chamber (plasticjar; 45-mm inner diameter; volume, 70 ml) containing a pieceof hybridization buffer-saturated tissue paper at the bottomof the chamber. The incubation chambers with the wells wereincubated for 2 h at 46°C. After incubation, 2 ml of a preheated(48°C) washing buffer (0.9 M NaCl, 20 mM Tris-HCl [pH 8],5 mM EDTA, 0.01% sodium dodecyl sulfate) was added to the wells,which were further incubated for 10 min at 48°C. The wellswere then placed back on the base of the filtration units, andtheir contents were washed by several cycles of addition andthen removal of distilled water. As before, special care wastaken not to expose the filter to the air. With approximately300 µl of the last wash still in the well, 500 µlof a DAPI (4',6'-diamidino-2-phenylindole) solution (1.5 µgml^-1; Sigma-Aldrich, St. Louis, Mo.) was added and incubatedin the dark for 5 min at room temperature. The DAPI solutionwas removed by washing the filter three times with 1 ml of distilledwater using the vacuum filtration unit as described above.

View larger version (63K):
[in this window]
[in a new window]
FIG. 1. Several essential steps for localization of target cells in the combined FISH-SEM method. The images show the Transwell filtration system used to collect and process cells (A), marking of the circular area with the target cell(s) in the middle (B and C), and cutting of the marked area under the dissecting microscope (D).

Preparing cells for SEM.
The probe-stained and DAPI-counterstained cells were dehydratedas follows. The distilled water remaining in the well afterDAPI staining was gradually replaced by washing the filter twotimes with 30% ethanol, using the vacuum filtration system asdescribed above. The wells were transferred to the tissue culturetray, a fresh portion (2 ml) of 30% ethanol was added, and thewells were left in the dark for 12 min. This process was sequentiallyrepeated with 50, 70, 85, 95, and 100% ethanol (the last stepwas repeated three times for 15 min each). As before, it wasessential that the filter not be exposed to air at any pointduring the procedure. After the last portion of 100% ethanolwas removed by use of a vacuum (except for a small amount thatwas just enough to cover the membrane), the wells were quicklyplaced into a tissue culture tray, covered with 2 ml of a mixtureof 100% ethanol-hexamethyldisilazane (Electron Microscopy Science,Ft. Washington, Pa.) (1:1 [vol/vol]), and allowed to be infiltratedfor 20 min in the dark. This was followed by two 20-min infiltrationsteps with 100% hexamethyldisilazane. At this point, cells wereallowed to air dry, as this caused no significant cell distortion.In this state, the filter can be stored in a dark and dry placeat room temperature for at least 1 week with no losses in fluorescenceintensity.

Fluorescence microscopy.
The air-dried filters were cut out of the wells with a dissectingknife and mounted on an adhesive silicone spacer (Schleicher& Schuell MicroScience, Riviera Beach, Fla.) affixed toa glass microscope slide. The filters were scanned under appropriateepifluorescent illumination on Zeiss Axioplan 2MOT and Axioskop50 compound microscopes equipped with an HBO 100-W mercury lamp;x10 Neofluar, x40 (dry) Neofluar, x40 (oil) F-Fluar, and x100Apo objectives; DAPI- and Cy-3-specific filter sets; and anHitachi ORCA cooled charge-coupled device camera (Hamamatsu,Hamamatsu City, Japan) operated by the OpenLab software package(Improvision Inc., Lexington, Mass.). Once the positively Cy-3-stainedtarget cell was located by use of the x10 and x40 (dry) objectives,a filter area of approximately 5 mm in diameter around the cellwas marked with a needle; these and other steps are illustratedin Fig. 1. The marked area was cut out under a dissection microscope(Zeiss Stemi 2000-C) with a dissecting knife. The cut-out piecewas checked once again under the fluorescence microscope (x10and x40 dry objectives) to confirm the presence and locationof the target cell(s). At this point, a digital photograph ofthe target cell(s) and surrounding areas was taken, with specialcare taken to capture small landmarks (atypical cell aggregations,unusually shaped particles, etc.). This facilitated locatingthe cell under SEM. The cut-out piece was then attached to acarbon adhesive tab, mounted on an SEM specimen holder, andsputter coated with 10 to 15 nm of gold-palladium (60:40) byuse of a Tousimis Samsputter 2A. SEM was performed on an AmrayAMR-1000 scanning electron microscope. The target cells werelocated at a low SEM magnification, using photographs takenwith the epifluorescence microscope.

Some filters were used to examine the Cy-3-labeled cells byepifluorescence rather than SEM. These were mounted on a glassslide, a mixture of 1 part VectaShield (Vector Laboratories,Burlingame, Calif.) and 4 parts Citifluor AF1 (Citifluor, London,United Kingdom) (vol:vol) was added, and a coverslip was gentlyapplied at an angle. This allowed the use of high-numerical-aperture(NA) lenses (x40 oil and x100 oil) to obtain high-quality epifluorescenceimages of the target cells. Throughout the project, the principalcharge-coupled device camera parameters remained constant (theexposure time was 0.2 s, and all other parameters were at theirdefault settings), except in control preparations, in whichcases the exposure time was increased to 2.0 s to avoid featurelessblank photos and achieve at least some visualization of thecontrol specimens.

Target cell recovery.
We evaluated the usefulness of our FISH protocol to detect andquantify the target cells in complex species mixes and in environmentalsamples. The species mixes were prepared using the strains listedin Table 1. Known numbers of Euplotes sp. cells were added tovials with 3 ml of artificial protistan mixes to achieve concentrationsof 2, 4, and 8 Euplotes sp. cells/ml. The contents of five vialsfor each Euplotes sp. concentration were fixed and hybridizedto Euplotes-specific probes as well as to the universal andnonsense probes as described above. The negative control containedno probe. All filters were counterstained with DAPI (see above)to obtain the total protistan count. The resulting filters werescanned to examine the specific and nonspecific staining oftarget and nontarget cells and to count the Cy-3-labeled Euplotesspp.

In addition, 10 Euplotes sp. cells were added to four samples(10 ml each) obtained from Sippewisset salt marsh, which weresubsequently stained with the two Euplotes-specific probes.Controls included environmental samples with naturally occurringEuplotes spp. only. For each probe and control treatment, duplicatesamples were processed as described above, and duplicate filterswere prepared and scanned to detect the Cy-3-labeled targetand nontarget cells.

RESULTS AND DISCUSSION

Top

Results and Discussion

Fixation.
The quality of fixation emerged as one of the most importantfactors determining the quality of SEM images. The majorityof trials resulted in collapse of most cells and clearly substandardSEM images (a typical result is illustrated in Fig. 2). We examinedhow the cell morphology was changed with each step of the FISHprotocol and found that most damage occurred during the dehydrationprocedure. We addressed this both by modifying the way the cellswere handled during FISH and by identifying the fixative thatprovided sufficient preservation of the cells. Of the nine differentfixatives and their combinations tested, only the Bouin's fixativemixture supplemented with glutaraldehyde imparted enough rigidityto the cells such that they retained their shape during FISHand SEM. The qualitatively best combination of cell rigidity,strength of specific fluorescence signal, and low nonspecificbackground fluorescence was achieved at 50% Bouin's fixative(full-strength solution mixed with sample of equal volume) supplementedwith 0.1 to 0.2% glutaraldehyde (see Fig. 3 to 5). Interestingly,Bouin's fixative mixture (at a different concentration and withno glutaraldehyde added) was also found to be optimal for combiningFISH and silver staining protocols for ciliate identification(7).

View larger version (137K):
[in this window]
[in a new window]
FIG. 2. Typical SEM images of collapsed Euplotes eurystomus. Heavy damage to the cell morphology was observed with all fixatives except Bouin's fixative supplemented with glutaraldehyde.

View larger version (68K):
[in this window]
[in a new window]
FIG. 3. Simultaneous FISH and DAPI staining of several test microeukaryotes in mixed samples. (a to e) Universal eukaryotic probe Euk1209R/DAPI. (f to h) Negative controls with no probe or DAPI (f and g) and with a nonsense probe and DAPI (h). (a) Euplotes eurystomus; (b) Prorocentrum micans and Euglena gracilis (note the lack of probe binding to the latter); (c) Aspidisca sp.; (d) Colpidium striatum; (e) Dileptus cygnus; (f) Euplotes eurystomus; (g) Ochromonas danica; (h) diverse organisms in mixed sample. Bar = 10 µm.

View larger version (80K):
[in this window]
[in a new window]
FIG. 5. SEM images of organisms retrieved from mixed cultures by use of the combined FISH-SEM protocol. (a) Chilomonas paramecium; (b) Colpidium striatum; (c) Dileptus cygnus; (d) Euglena gracilis; (e) Euplotes eurystomus (ventral surface); (f) Euplotes eurystomus (dorsal surface); (g) Peridinium cinctum; (h) Prorocentrum micans; (i) Synura uvella.

FISH staining.
The commonly used FISH protocol (16) includes a filter-dryingstep, which serves two main purposes. First, it is a means toattach the target cells to the membrane and prevent excessivecell losses during washing steps. Second, it enhances probepenetration into the target sites in the cell (ribosomes). Inour observations, most wall-less organisms collapsed duringair drying and became largely unrecognizable. This called fora modification of the standard FISH protocols. We discoveredthat the cell collapse could be easily prevented if the targetcells were kept in solution at all times during FISH. To minimizecell losses, we collected cells in plastic wells with a polycarbonatemembrane bottom, as described in a previous protocol (6). Whensuch a well is placed on the base of a typical filtration unitconnected to a pump (Fig. 1), reagents can be added to and removedfrom the well in a controlled fashion. Since the cells stayinside the well during the entire procedure, their losses becomeminimal (6) (also see below). After the FISH protocol is completed,the membranes can be cut out of the well and used for epifluorescenceobservations similarly to the standard protocols.

Both Euplotes sp.-specific probes provided satisfactory detectionof this ciliate (Fig. 3 and 4). After we optimized one of theprincipal FISH protocol parameters (i.e., concentration of formamide;data from preliminary trials are not shown), the specificityof staining became 100%. Neither of the two Euplotes sp.-specificprobes stained any of the nontarget cells. The nonsense probedid not stain any of the DAPI-counterstained protistan cells,and the level of background fluorescence was sufficiently low(Fig. 3). The universal probe hybridized to most but not allmicroeukaryotes, which indicated that there were mismatchesin the target sequences (SSU rRNA gene) for at least some protists(Fig. 3). Overall, the fluorescence signal was stable for atleast 1 week for sample storage in the dark at room temperature.

View larger version (71K):
[in this window]
[in a new window]
FIG. 4. Images obtained at different stages of the combined FISH-SEM protocol. (a) Euplotes eurystomus, Paramecium caudatum, and Colpidium striatum retrieved from a culture mix (10 strains) by use of probe Euk1209R. (b) Euplotes spp. retrieved from an environmental sample by use of probe Eupl240. Top panels, DAPI staining in UV channel; second panels from top, FISH staining in the red channel; third and bottom panels, the same areas under SEM.

We concluded that the modified FISH protocol used was a robusttool for visualization of target cells with an epifluorescencemicroscope at mid-range magnification (x40 dry objective).

We checked the degree of cell recovery by the modified FISHprotocol. We added a known number of Euplotes sp. cells to environmentalsamples, quantified the increase in the Euplotes sp. count,and compared this increase with the number of added cells (Table2). The recovery rate was 92% ± 21% for the Euplotes-240probe and 112% ± 21% for the Euplotes-1780 probe. Wealso attempted to quantify the recovery of Euplotes spp. froma mix of 11 different protistan cultures (species are indicatedin Table 1). The recovery values (for three independent trials)ranged from 93% ± 9% to 97% ± 5%.

View this table:
[in this window]
[in a new window]
TABLE 2. FISH-aided recovery of Euplotes sp. cells added to environmental samples containing naturally occurring Euplotes

Bridging FISH and SEM.
Once the target cell is located on the filter by use of theepifluorescence microscope, its position has to be registeredto facilitate the search for the same individual cell underSEM. We developed a simple approach to keep track of the cellsof interest. This technique consists of several steps: centeringthe target (stained) cell in the microscopic field of view ofthe epifluorescence microscope, turning on the transmitted lightsource to create a light circle around the cell, closing thefield diaphragm to bring the diameter of the circle to approximately3 to 5 mm (Fig. 1B), dotting or puncturing the membrane alongthe edges of the illuminated circle (Fig. 1C), and cutting asmall piece of the membrane containing the circle and the cellof interest in its center (the last step was performed undera dissecting microscope; Fig. 1D). The resulting piece of themembrane can be easily mounted on an SEM specimen holder andscanned under SEM at low magnification to locate the targetcell.

SEM.
Once the details of fixation, FISH, and cell handling were workedout, preparation for SEM and SEM observations were done accordingto standard protocols. We experienced no difficulties in locatingthe target cells by SEM (Fig. 4), and the quality of materialobserved was high (Fig. 5). We used a wide selection of organismsacross the Protista (cryptomonads, kinetoplastids, stramenopiles,and alveolates) and found all the test species exhibiting theirtypical morphological characteristics, including those thatare important taxonomically (Fig. 5). The images obtained fromthese preparations are comparable to those obtained by routineSEM techniques alone. The ciliates Dileptus cygnus and Colpidiumstriatum retained their characteristic shapes and morphologicaldetails (Fig. 5b). Euplotes spp. exhibited well-preserved frontoventral,transverse, and caudal cirri as well as the adoral zone of membranelles(Fig. 5e and f). The dinoflagellate Prorocentrum micans (Fig.5h) showed the expected bilateral compression and two thecalplates typical of the genus (and the order Prorocentrales ingeneral). Images of dinoflagellate Peridinium cinctum revealedthis organism's diagnostic tabulation pattern and pronouncedornamentation of thecal plates (Fig. 5g). Representative cellshape, helical striations, and position of the single flagellumwere exhibited by the kinetoplastid Euglena gracilis (Fig. 5d).Stramenopile Synura petersenii was seen with its distinctivesilica scales (Fig. 5i), and Chilomonas paramecium displayedthe normal cryptomonad cell architecture (Fig. 5a). We concludedthat the method developed allowed visualization of importantdiagnostic morphological characteristics across a wide rangeof protistan diversity and is therefore an adequate tool forstudying the morphologies of novel organisms.

Application in the field.
One specific aim of this research was to ensure that our FISH-SEMapproach could be used in field-oriented research. In the field,the use of FISH is often impractical or impossible, and so someform of prolonged cell storage must be found. We discoveredthat Bouin's fixative-glutaraldehyde-fixed cells inside wellscould be stored for at least several weeks at -20°C if coveredwith a mix of 3 parts 1x PBS and 2 parts 100% ethanol. The wellsshould be placed into tissue culture trays and sealed with Parafilm.Such storage did not appear to compromise the ability of cellsto hybridize well with FISH probes.

Conclusions.
We have developed a methodological approach that combines FISHand SEM for studies of protists. It consists of original andmodified protocols of standard fixation, FISH, and SEM. Theindividual components of this approach work together seamlesslyand are compatible with the demands of field-based research.The approach provides high-quality SEM images of target cellsthat are first identified by FISH. It may therefore serve asa convenient tool to conduct ultrastructural studies of organismsknown exclusively from their 18S rRNA signatures.

ACKNOWLEDGMENTS

We thank Klaus Hausmann (University of Berlin), Denis Lynn (Universityof Guelph), and Eugene Small (University of Maryland) for theiradvice on SEM and Christine Beardsley (MPI for Marine Microbiology,Bremen, Germany), Johannes Fried (Technical University, Munich,Germany), and Michael Schweikert (Technical University, Stuttgart,Germany) for helpful discussions on FISH. Ed Jarroll (NortheasternUniversity, Boston, Mass.) was instrumental in providing theinstitutional support for this project.

This work was funded by Deutsche Forschungsgemeinschaft grantSTO414/2-1 to T.S. and by U.S. National Science Foundation grantsto S.S.E. (OCE-9618135 and DEB-0103599).

FOOTNOTES

* Corresponding author. Mailing address: Northeastern University, Marine Science Center, East Point, Nahant, MA 01908. Phone: (617) 373-4048. Fax: (617) 373-3724. E-mail: s.epstein{at}neu.edu.

This is contribution 248 of the Marine Science Center of NortheasternUniversity, Nahant, Mass.

REFERENCES

Top