Cytometry 11:132-143 (1990)

1990 W d e y - L m , Inc.

Flow Cytometric Detection of p-Globin mRNA in Murine Haemopoietic Tissues Using Fluorescent In- Situ Hybridizationr J a n A. Bayer' and J a n G.J. Bauman TNO Radiobiological Institute, Rijswijk (J.A.B., J.G.J.B.); Erasmus University, Rotterdam, The Netherlands (J.G.J.B.)

The novel method for flow cytometric detection of cellular RNA species in suspended cells by fluorescent in situ hybridization (FC-FISH) was applied in the evaluation of p-globin expression in murine haemopoietic tissues. Normal murine bone marrow cells and regenerating bone marrow cells obtained after lethal irradiation and bone marrow transplantation as well as murine 15 d fetal liver were examined. Furthermore, spleens and bone marrow of phenylhydrazine-induced anaemic mice were studied. Biotinylated sense- and antisense single strand RNA probes, obtained by transcription of a 510 nucleotides murine p-globin cDNA sequence subcloned into the pGEMl plasmid were used as hybridization probes. For detection of the hybrids formed, avidin-FITC was used. Only the antisense 0-globin probe gave strongly positive fluorescence signals in a defined population of cells in each of the tissues examined, whereas the sense probe did not give signals higher than control samples. Melting characteristics of the hybrids showed the specificity of the in situ hybridization reaction. Forward light scatter distributions, reflect-

ing cell size of the positive cells were as expected from erythroid cells. Within the erythrocyte subpopulation both p-globin-negative and -positive cells were detected. The percentages of positive cells determined flow cytometrically correlated with the percentages observed' in May-Griinwald/Giemsa stained preparations. Differences observed in Buorescence intensity between positive cells of different organs were no larger than about a factor of two, indicating a rather constant p-globin mRNA content over the entire differentiation range. An exception was 15 d fetal liver, which was shown biochemically to contain about eight times more p-globin RNA and which had a 2.4 times higher fluorescence intensity. We estimate that the sensitivity of the present method is such that as little as 500 copies per cell of a specific mRNA of 1 kb length would be detectable. Key terms: p-globin gene expression, in suspension hybridization, single cell mRNA detection, biotinylated RNA probe, antisense RNA probe, bone marrow, fetal liver, anaemia

Flow cytometric analysis of fluorescent in situ hyNon-radioactive in situ hybridization methods have been rapidly emerging with t h e development of DNA- bridization (FC-FISH) results would allow for differenand RNA-haptenization methods. The enzymatic intro- tial quantitation of target nucleic acid contents of induction of biotin moieties (19) h a s had t h e most wide- dividual cells in large and heterogeneous populations. spread application, but also N-Acetoxy-N-Acetylaminofluorene (AAFbmodification (20,26), mercuration (5,151, and sulfonation (30) are being used more and more. In all of the labeling methods mentioned, post hybridization (immunob cytochemical detection is 'This investigation was supported by the KWF, Netherlands Canused. The main advantages of these non-radioactive cer Foundation, research project IKR 85-9. methods are the rapid availability of results, and the "Address reprint requests to J.A. Bayer, Department of Cytometry high spatial resolution when fluorescent tagging is ap- and Cytocheinistry, TNO Radiobiological Institute, Lange Kleiweg 151, 2288 GJ, Rijswijk, The Netherlands. plied.

Only a limited number of FC-FISH applications have been described to date. These encompass the detection of repetitive DNA-species in isolated nuclei (28,291 or suspended chromosomes (131,and the detection of high copy number ribonucleic acid species (4). Development of successful applications has been hampered by the incompatibility of hybridization conditions with the necessary preservation of cellular morphology (or a t least: particle integrity), as required in flow cytometry

generating bone marrow was harvested on d 7 after transplantation as indicated above. Fixation and storage. Single cell suspensions were fixed by adding a n equal volume of 2% formaliniHH and incubation for 15 min at room temperature. After washing with HH, cells were resuspended in 70% ethanol, washed once, and stored in 70% ethanol a t -20°C for up to 12 months.

(27).

Isolation of Total Cellular RNA Total cellular RNA was isolated from 3 x lo8 fetal liver cells and 3 x lo8 normal bone marrow cells by differential precipitation with 4M LiCl(11. Yields were determined spectrophotometrically, and 1 p1 volumes of twofold serial dilution series were spotted onto nitrocellulose, for filter spot hybridization.

In this paper we describe the use of FC-FISH to fixed cells in suspension (4) to measure P-globin mRNA contents. The cellular morphology was well retained after the FC-FISH procedure, and subsequent microscopical examination unequivocally indicated positive hybridization signals in cells of the erythroid lineage from different murine haemopoietic tissues.

MATERIALS AND METHODS: Preparation and Fixation of Single Cell Suspensions Male BCBA (C57BliRij x CBAIRij) mice were used throughout this study. The mice were bred and maintained under specific pathogen free conditions in the Radiobiological Institute TNO, until initiation of the experiments. Anaemic spleen (AS)and anaemic bone marrow (ABM). Mice received phenylhydrazine (20 mgiml in HEPES buffered Hank's Balanced Salts Solution without phenol red, osmolarity 300 ? 5 mOsm; HH, GIBCO, Santa Clara, CA) 100 mg/kg intraperitoneally (i.p.),on d 1 and d 3 . On d 5, spleens were excised after cervical dislocation. Single cell suspensions were prepared in HH by dispersion of whole spleens through nylon mesh. Bone marrow was harvested by flushing femora with approximately 1.0 ml HH. The cells were syringed gently with a 25-g needle and passed through nylon mesh in order to disperse cell clumps, and washed once with HH by centrifugation (10 min, SOOg, 4°C) and resuspended in HH. Fetal liver (FL). Fetal liver cell suspensions were obtained from mouse embryos, 15 d after gestation (timed matings). Pregnant mice were killed by cervical dislocation under C 0 2 anaesthesia, and uteri were removed and placed in HH. The embryos were dissected free from the uterus, and fetal livers were removed by use of a binocular dissecting microscope. Pooled fetal livers were disaggregated by gentle pipetting and passing through nylon mesh. Normal bone marrow (NBM) and normal spleen (NS). Normal bone marrow and spleen was harvested from 7 to 8 weeks old mice a s indicated above for anaemic bone marrow and spleen. Regenerating bone marrow (RBM). Bone marrow transplantation was accomplished by intravenous injection of 2.5 x 10' normal bone marrow cells into the lateral tail vein of lethally irradiated (9.25 Gy y total body irradiation) 12 weeks old BCBA male mice. Re-

Probes The recombinant plasmid pMPG(1.7) was constructed by subcloning the 1,700 nt Hha I fragment containing the 510 nt murine (3-globin cDNA from pCR1:pMS (24) into the HincII site of pGEMl (Promega Biotec, Madison, WI). Linearization before transcription was performed by using either EcoRI or Hind111 (GIBCO-BRL, GIBCO Europe, Breda, The Netherlands). In all of the experiments, orientationspecific single strand RNA probes were used that were generated by using SP6 o r T7 RNA-polymerase (GIBCO-BRL1 in in-vitro transcription assays, a s described earlier (4). In Situ Hybridization in Suspension In situ hybridization to suspended cells was performed essentially as described earlier (4).In short, fixed cells in 70% ethanol were DEPC-treated (0.2%,15 min a t room temperature), rehydrated in HH, and suspended in 5 x SSC/50%formamide. Hybridization was carried out for 4h in the presence of about 25 pgiml biotinylated single strand RNA probe. Stringency washing was performed in 0.5 x SSC/O.lR SDS a t 45°C (unless otherwise stated) for 30 min with constant agitation. After detection of the hybrids by using fluoresceinated avidin DCS (avidin FITC; Vector Laboratories, Burlingame, CA) cells were resuspended in HHI 0.05% Tween-20 (HH-T) containing 5 mM EDTA and 0.2 pg/ml4'-6-diamidino-2-phenylindole (DAPI; Sigma Chemical Co., St. Louis, MO). Filter spot hybridizations using biotinylated single strand RNA probes were carried out by using standard molecular biological techniques (22). In short, after prehybridization for 2 h a t 45°C in 50% formamid, 5 x SSC, 0.5 mgiml E . coli tRNA, 0.1% SDS, probe was added at about 1 pg/ml. Hybridization was performed overnight a t 45"C, and filters were washed two times for 15 min with 2 x SSC a t 45°C. The stringency wash for the filter hybridizations was the same as for the in suspension hybridizations. Filters were developed with streptavidin-alkaline phosphatase (GIBCO-BRL) and

134

BAYER AND BAUMAN

BCIPiNBT according to manufacturers' protocols to yield purple precipitates.

Flow Cytometry and Data Evaluation Fluorescence and light scatter intensities were measured by using the two-laser Rijswijk Experimental Light-Activated Cell Sorter (RELACS 111).A Coherent Innova 90-5 Argon-ion laser (Coherent, Palo Alto, CA), emitting 600 mW of power tuned a t 488 nm, was used to excite both the FITC and autofluorescence. Forward light scatter (FLS) and perpendicular light scatter (PLS) were also 488 nm derived. A second Argon-ion laser (Spectra-Physics 202015; Spectra-Physics, Mountain View, CA) tuned to emit 100 mW of UV (3511364 nm) was used to excite DAPI. Fluorescent light emission was spectrally separated by using the following filters: 530 nm bandpass (PlOi 530, Corion, Hollington, MA) for FITC, two 610 nm longpass (RG610, Schott Glass Technologies Inc., Duryea, PA) for autofluorescence, and two KV408 (Schott Glass Technologies Inc., Duryea, PA) plus one 450 nm shortpass (LS450, Corion) filter for DAPI fluorescence. Perpendicular light scatter signals were detected by using a 488 nm bandpass filter (03FIL002, Melles Griot, Irvine, CA). Pulse width (time of flight, TOF) of the triggering signal (FLS) was routinely measured. Logarithmically amplified (FITC and autofluorescence) or linearly amplified (FLS, PLS, DAPI, and TOF) signals were stored as six-parameter list mode data sets on a Hewlett Packard 9000-220 microcomputer. Data analysis was performed by using the Eightparameter Listmode Data Analysis Software (ELDAS), as developed in the Radiobiological Institute TNO (17). Mean fluorescence intensity (f.i.1 was calculated from logarithmic distributions after determining that a 60-channel increment on a logarithmic scale is a factor of ten on a linear scale. Channel 1 on the logarithmic scale was taken to be 1.0 arbitrary unit on the linear scale. The equation used to calculate f.i. values was:

log FITC fluorescence FIG 1. FITC fluorescence profiles of normal murine bone marrow cells after FISH in suspension. Cells were hybridized in suspension with P-globin antisense probe (heavy line), sense probe (thin line), or without addition of probe (dotted line). From list mode data, singlets (90Q of total) were selected by gating on TOF and DAPI fluorescence, and the respective FITC fluorescence histograms are shown. FITC fluorescence intensity is shown on a "log-scale (60 channels is one decade).

ange-red) from FITC fluorescence (green-yellow). Exposures were timed with a Photomat automatic exposure timer. For color slides Kodak Ektachrome P8001 1600 was used and processed a s 1,600 ASA.

RESULTS Iletection of P-Globin mRNA and Hybrid Stability Normal bone marrow cells were hybridized in suspension with either sense- or antisense probe, or without addition of probe. After washing and staining with avidin-FI'I'C and DAPI, cells were measured on the RELACS-111. In Figure 1the FITC-fluorescence distributions are shown that were obtained. Cells hybridized IJ 60. n, with the antisense probe, which is complementary to f.i.=C r=I, N mature p-globin mRNA in the cells, give fluorescence In this equation, L and U are the lower and upper histograms in which three subpopulations can be diswindow boundaries that were used for selection of the cerned (Fiig. 1,thick line). Strong positive cells (29% of specific population, z is the channel number (0-255), ni all cells; mean fluorescence intensity (f.i.) : 77), cells is the number of events in channel i, and N is the total showing dull (auto)fluorescence (54%; f.i.: 241, and real number of events in window L-U. negative cells (17%, f.i.: 6). The negative cells predominantly consisted of anucleate erythroid cells (> 85%), Microscopy whereas the dull cells were nucleate cells that had acFluorescence microscopy was performed to confirm quired au tofluorescence during the hybridization prohybridization results after flow cytometry . Remaining cedure. No-probe controls and cells hybridized with cells were collected by using centrifugation (10 min, sense probe showed only two peaks, i.e., the dull auto1,500g) and resuspended in antifade (50 g/L 1,4-diazo- fluorescent cells (61% and 49%; f.i.: 30 and 31, respecbicyclo-[2,2,21octane in glycerol/ 0.1 M Tris-HC1, pH tively) and the negative cells (39% and 52%; f.i.: 6 and 8.0 (90:10 viv; ref. 8). Observations and photography 7, respectively). It is clear that the sense probe (the were performed on a Leitz Dialux fluorescence micro- same orientation as the mature P-globin mRNA, therescope with omission of the red-block filter routinely fore non-hybridizing, and serving a s a control for nonused for FITC fluorescence observations. This was done specific sticking of probe) does not give rise to a marked in order to be able to distinguish autofluorescence (or- increase in fluorescence, indicating that the stringent

FLOW CYTOMETRIC DETECTION O F p-GI,OBIN m K N A

.-c v) c

a

75

-

50

-

25

-

c

c ._

a

0 K

a

0 v)

2

0

3 'c

LL

135

Total Cellular RNA Spotblot-Analysis The P-globin mRNA content of RNA isolated from normal bone marrow and fetal liver cells was determined as follows. Equal amounts of total cellular RNA were spotted in twofold dilution series on a nitrocellulose filter and hybridized with sense and antisense biotinylated p-globin probes. The results of detection of hybrids with streptavidin-alkaline phosphatase and developing with BCIPiNBT are shown in Figure 3. For fetal liver the 1ng spot is clearly visible and has about the same intensity a s the 8 ng spot in the normal bone marrow lane. This indicates that the p-globin mRNA content of fetal liver RNA is about eight times higher than the p-globin mRNA content of normal bone marrow RNA. The amount of RNA isolated from 3x10' normal bone marrow and fetal liver cells was about 0.5 mg and 1 mg, respectively. For control purposes, next to the cellular RNA's non-biotinylated sense and antisense transcripts a s well as denatured, EcoRI-linearized pMPG(1.7) was spotted onto the filters. Results indicate that sense probe hybridizes to antisense transcript, and inversely antisense probe hybridizes to sense transcript with approximately equal efficiency (Fig. 3D,E). Both probes hybridize to the p-globin insert harbouring plasmid (Fig. 3F).

1 0

40

50

60

70

80

90

100

temperature ("C) FIG. 2. Temperature dependency of p-globin antisense probe derived FITC fluorescence intensity Ki.1 after FISH in suspension. Murine 7 d regenerating bone marrow cells were hybridized with pglobin antisense probe, and washed a t increasing temperature in 0.5 x SSC/O.l%~ SDS. DAPI-negative cells were selected from the data sets obtained, and f.i. values were calculated from the p-globin-negative subset (closed circles) and the B-globin-positive subset (open circles), and plotted.

Differential Analysis of P-Globin Expression in Murine Haemopoietic Tissues All haemopoietic tissue types, being NBM, RBM, ABM, NS, AS, and FL cell suspensions, were hybridized with sense and antisense biotinylated P-globin RNA probe. Bivariate distributions of forward light scatter vs. FITC-fluorescence and DAPI fluorescence vs. FITC fluorescence were constructed and are shown washing applied results in removal of all unhybridized in Figure 4. Univariate FITC-fluorescence distribuprobe. In all other hybridization experiments that are tions of the single-cell populations (obtained by exclubeing discussed, the curves for no-probe and sense sion of cell clumps using time of flight and DAPI fluprobe controls were virtually identical. Therefore only orescence as discriminative parameters) are also the respective sense probe control curves are given in shown (Fig. 4). From the list mode data and fluorescence histograms the percentages of p-globin positive the figures below. To determine the stability of the hybrids obtained we cells, a s well a s the net f.i. values (f.i. of the positive subjected regenerating bone marrow cells hybridized to cells in the subset minus f.i. of the negative cells in the sense or antisense p-globin probes to a range of increas- subset) were calculated and are given in Table 1. The ing temperatures. From the data sets obtained, DAPI- cell cycle status of the P-globin-positive and -negative negative cells were selected, and FITC-fluorescence cells was calculated from the DAPI histograms by usdistributions were generated from these subpopula- ing the rectangle method (Table 2; ref. 2). Furthertions. The f.i. values of the clearly separated positive more, differential counts were performed on Mayand negative subsets were calculated and plotted vs. Griinwald/Giemsa stained cytocentrifuge preparations temperature (Fig. 2). Melting of hybrids is nearly com- that had been prepared from fresh cells from the same plete a t a temperature of 95"C, and f.i. values almost samples that were used in the FC-FISH experiments reach "no-probe'' control levels. The temperature a t (Table 3). In the following, the combined results are which 50% of the signal was lost (at about 40 arbitrary being discussed for each of the haemopoietic tissues units in Fig. 2) was determined to be about 86°C. This separately. is in excellent agreement with the calculated melting Bone marrow. Comparison of overall FITCtemperature (T,) of RNA.RNA hybrids in 0.5 x SSC, fluorescence distributions of ABM and RBM with NBM which is 85°C for sequences with a G.C content of 50% (Fig. 4F,I,C, respectively) reveals a significant increase (as for the P-globin probe) and a length of 100-150 n t in the number of P-globin-positive cells in both ABM and RBM. This increase is mainly due to a strong in(9).

136

BAYER AND BAUMAN

A B C D E F

antisense

A B C D E F

sense

FIG.3. Filter spot hybridization with biotinylated sense and antisense p-Globin RNA probes of duplicate filters containing 1 (*I volumes of twofold dilution series of HL60 RNA (A),fetal liver RNA (B), normal bone marrow RNA ( C ) ,“sense” p-globin RNA (D),“antisense” p-globin RNA (E),and p-globin containing pGEM plasmid DNA (F).

The first dilution containes 500 ng of RNA (A-C), 4 ng of RNA transcript (D,E), or 100 pg of EcoRI-linearized pMpG(1.7) plasmid tF). After hybridization and washing (0.5 x SSC/O.l%SDS, 45”Ci, bound probe was visualized by using streptavidin-alkaline phosphatase and NBTIBCIP.

crease in the number of nucleate P-globin-positive cells (Table l),and it was largely confirmed in the differential counts (Table 3). Calculation of net f.i. of positive cells in the different samples indicated that p-globin mRNA content is somewhat higher in ABM and RBM a s compared to NBM (net f.i. 74 and 78 vs. 57, respectively). Bivariate dotplots of forward light scatter vs. FITCfluorescence illustrate that major differences exist in the composition of the tissues with respect to subpopulations expressing P-globin. In NBM, the majority of p-globin expressing cells are in the erythrocyte cluster, as defined on the basis of FLS and DAPI (Fig. 4A,B). In anaemic and regenerating bone marrow, many more large cells expressing P-globin are found, comprising nucleated erythrocyte precursors (Fig. 4D,E, and 4G,H). In all of the bone marrow samples, the percentage of @-globin-negative/DAPI-negativecells is in the order of 50% (Table 1).This is in concordance with the relative amounts of reticulocytes and erythrocytes in these samples, shown in Table 3. To show the morphological characteristics of pglobin-positive cells of NBM, forward light scatter vs. perpendicular light scatter dotplots were constructed of

all cells as well as of P-globin-positive cells. Also shown are forward light scatter histograms of these populations (Fig. 5). From this i t is evident that most of the @-globin-positivecells in NBM are in the erythrocyte cluster, a:: it is defined on the basis of FLS vs. PLS. Lower amounts of P-globin-positive cells however are found in the lymphocyte cluster (presumably normoblast) a s well as in the granulocyte cluster (presumably erythroblasts). The observed net f.i. of positive cells in the anucleate population was somewhat higher than in the populi3tion containing nucleated cells (Table 1).

FIG.4. Multivariate analysis of murine haemopoietic tissues. Cells were hybridized in suspension with biotinylated p-globin RNA probes, and hybrids were detected by using avidin-DCSiFITC. Total DNA was counterstained with DAPI. The bivariate scatterplots all contain 1,000 events from the respective list mode data sets, gated on DAPI fluorescence and TOF to exclude cell clumps and debris. Univariate FITC fluorescence distributions of the respective singlet cell populations, obtained when using antisense p-globin probe (thick lines) and sense p-globin probe (thin lines), are also shown. FLS and DAPI fluore:jcence are displayed on a linear scale (FLS amplifier gain is 2.5 times lower for FL than for the other samples); FITC fluorescence is shown on a ‘“log-scale (60 channels is one decade).

1

A

_._-. ......

I

r

&$@$&:;

*:::.. &. .... ... .;:

. . . ..

,

r

.. ;.-...., .. . . .. .. .. .. . . > * - . .. -.:' . . . -.:.: . ..... . i

-

FL

.I.-....... '. F LS

I DAPl

log FlTC fluorescence (a.u.) Fig. 4.

138

KAYER AND BAUMAN

Table 1 Analysis of p-Globin mRNA Containing Cells i n Different Murine Haenzopoietic Tissues by Using Flow Cytometric FISH Sample Normal bone marrow Anaemic bone marrow Regenerating bone marrow Normal spleen Anaemic spleen Fetal liver

% Nucleated"

% positive cells Nucleated Anucleate 16 44

All 57

Fluorescence intensityb Nucleated Anucleate 57 68

63

All 27

75

47

44

!56

74

70

82

41

53

51

.54

78

76

90

83 51 82

3 45 59

1 63 61

9 26 53

75 104 138

67 97 106

81 110 182

"Percentages nucleated cells were determined from list mode data sets cm the basis of DAPI fluorescence after exclusion of cell clumps and debris. 'The net f.i. values of the positive subpopulations were determined by subtracting the f.i. values of the negative cells in the same subpopulation from the f.i. values of the positive cells. Negative cells' f.i. values ranged from 29 to 34 for nucleated cells and from 6 to forty for anucleate cells (about the same level as control probe levels). Table 2 S-phase Content (%I of the Nucleated Cells Subpopulation of Different Haemopoietic Tissues i n Relation to P-Globin Expression" Sampleb Normal bone marrow Anaemic bone marrow Regenerating bone marrow Normal spleen Anaemic spleen Fetal liver

All 13

% S-uhase Positive cells Negative cells 36 8

24

40

12

29

43

17

2 33 36

21 43 41

19 32

1

"The relative volume of the S-phase compartment was estimated by using the rectangle method (2) and is expressed as a percentage of all nucleated cells after exclusion of cell clumps and debris. bCV values of the GI,,-peak in the different samples ranged from 0.020 to 0.053 (average 0.033).

Spleen. In normal spleen, only very minor numbers of positive cells are detected (Table 1;Fig. 4L). Most of those are DAPI negative (Fig. 4J,K). Phenylhydrazine treatment results in a strong increase in the number of P-globin-positive cells (Fig. 4M-0). Of all nucleate cells, 63% are P-globin positive in anaemic spleen (Table 11, nearly the same percentage that was found for erythroid precursors in differential counting (Table 3). From these erythroid precursors, the vast majority (over 60%) were found to be polychromatic normoblasts (Table 3). Calculated overall net f.i. values were significantly higher for AS than for NS (104 and 75, respectively), and were higher in the anucleate subpopulation (110 and 81, respectively). As can be seen from the DAPI vs. probe fluorescence plot (Fig. 4N), P-globin mRNA containing cells are actively proliferating. From Table 2 it can be derived however that the Sphase content of the P-globin mRNA containing subpopulation is not markedly elevated in any of the hae-

mopoietii: tissues examined. This might indicate that cell cycle time of erythroid precursors is not strongly affected in the anaemic or regenerating status. Only the total number of P-globin-positive cells seems to be modulated (Table l),in parallel with a n increase in the number of erythroid precursors observed in differential counting (Table 3). Fetal liver. Differential counts of 15 d post gestational fetal liver indicate that almost all cells (94%)are of the erythroid lineage (table 3). These cells are quite large, as is also indicated by the fact that amplification of the FI,S signals depicted in Figure 4P was 2.5 times decreased as compared with all of the other samples. Careful examination of FITC fluorescence distributions reveals the presence of a population of brightly fluoresceing cells and another population of intermediately positive cells (Fig. 4R). The cluster of intermediately positive cells is clearly isolated in the FLS vs. FITC fluorescence plot (Fig. 4P), and contains 18%of the cell!$ (f.i. 95). This cluster consists of very large cells, presumably erythroblasts, containing lower amounts of P-globin mRNA. The strongly positive cells are smaller and contain larger amounts of P-globin mRNA (42%; f.i. 202). These are probably the more mature erythroid series cells. A lower level threshold on FITG fluorescence was set in such a way that

Flow cytometric detection of beta-globin mRNA in murine haemopoietic tissues using fluorescent in situ hybridization.

The novel method for flow cytometric detection of cellular RNA species in suspended cells by fluorescent in situ hybridization (FC-FISH) was applied i...
1MB Sizes 0 Downloads 0 Views