Abstract
Fluorescence in situ hybridization (FISH) is a powerful means by which single- and low-copy DNA sequences can be localized on chromosomes. Compared to the mitotic metaphase chromosomes that are normally used in FISH, synaptonemal complex (SC) spreads (hypotonically spread pachytene chromosomes) have several advantages. SC spreads (1) are comparatively free of debris that can interfere with probe penetration, (2) have relatively decondensed chromatin that is highly accessible to probes, and (3) are about ten times longer than their metaphase counterparts, which permits FISH mapping at higher resolution. To investigate the use of plant SC spreads as substrates for single-copy FISH, we probed spreads of tomato SCs with two single-copy sequences and one low-copy sequence (ca. 14 kb each) that are associated with restriction fragment length polymorphism (RFLP) markers on SC 11. Individual SCs were identified on the basis of relative length, arm ratio, and differential staining patterns after combined propidium iodide (PI) and 4′,6-diamidino-2-phenylindole (DAPI) staining. In this first report of single-copy FISH to SC spreads, the probe sequences were unambiguously mapped on the long arm of tomato SC 11. Coupled with data from earlier studies, we determined the distance in micrometers, the number of base pairs, and the rates of crossing over between these three FISH markers. We also observed that the order of two of the FISH markers is reversed in relation to their order on the molecular linkage map. SC-FISH mapping permits superimposition of markers from molecular linkage maps directly on pachytene chromosomes and thereby contributes to our understanding of the relationship between chromosome structure, gene activity, and recombination.
FOR most plant species, what is known about the order of loci on chromosomes is based almost entirely on genetic linkage maps. Such linkage maps are generated by producing multi-hybrid crosses and determining the relative frequency of recombination between genes or molecular markers (Tanksleyet al. 1989; Paterson 1996). The physical location of genes on chromosomes is of great interest because chromosome structure profoundly influences gene activity (see Lohe and Hilliker 1995; Wallrath and Elgin 1995; Zuckerkandl and Hennig 1995 for reviews). However, linkage maps cannot be superimposed on chromosomes because map distances are not proportional to physical distances (Sturtevant and Beadle 1939; Khush and Rick 1968; Moore and Sherman 1974; Flavellet al. 1985). The relative scarcity of genes and crossing over in heterochromatin is partly responsible for this discrepancy, but crossing over is not evenly distributed in euchromatin either (Snapeet al. 1985; Dooner 1986; Tsujimoto and Noda 1990; Curtis and Lukaszewski 1991; Kotaet al. 1993; Stacket al. 1993; Sherman and Stack 1995). The most direct means of determining the location of genes on chromosomes is through the use of fluorescence in situ hybridization (FISH). In this technique hapten-labeled DNA probes are hybridized to chromosomes that have been spread on glass microscope slides, and antibodies or other affinity reagents conjugated to fluorochromes are used to detect directly or indirectly sites of hybridization (Trask 1991). FISH with repetitive sequences as probes has been widely reported for both animal and plant chromosomes, and single-copy FISH to mammalian chromosomes is fairly routine (e.g., Landegentet al. 1987; Lawrenceet al. 1988; Lichteret al. 1991; Henget al. 1992). Because most techniques used to prepare plant chromosomes leave overlying debris that interferes with probe penetration, single-copy FISH to plant chromosomes has proven considerably more difficult (Lehferet al. 1993; Jianget al. 1995). However, now there are a few laboratories that have overcome this obstacle and are successfully localizing single-copy sequences on plant chromosomes by FISH (e.g., Leitch and Heslop-Harrison 1993; Hansonet al. 1995; Jianget al. 1995; Gómezet al. 1997).
FISH is usually performed on mitotic metaphase chromosomes, but there are reasons to believe that pachytene (meiotic) chromosomes may be better substrates. Each pachytene chromosome (bivalent) is composed of two homologous chromosomes that are joined along their entire length by a proteinaceous scaffold called the synaptonemal complex (SC; Moses 1968). Because each homologue contains two chromatids, there are four closely associated copies of each locus available for hybridization on a bivalent. In comparison, there are only two nearby copies of each locus available for FISH on a metaphase chromosome. In spreads of pachytene chromosomes that have been prepared to reveal SCs (SC spreads), chromatin extends as a diffuse cloud around each SC. The loops of DNA extending from the SC (e.g., Weith and Traut 1980) appear to be more accessible to FISH probes than the DNA of condensed metaphase chromosomes (Moens and Pearlman 1989, 1990; Henget al. 1994; Solari and Dresser 1995), and SC spreads can be prepared relatively free of overlying debris. Additionally, pachytene chromosomes are 5–15 times longer than corresponding metaphase chromosomes (Ramanna and Prakken 1967; Stack 1984), so many closely associated loci that are not resolvable by FISH on metaphase chromosomes should be resolvable on pachytene chromosomes.
Here we report high resolution localization of two single-copy sequences and one low-copy sequence on tomato SC 11 using FISH. This is the first report of single-copy FISH to SC spreads and one of only a few studies in which FISH has been used to study the relationship between genetic linkage and chromosome morphology in plants (e.g., Pedersen and Linde-Laursen 1995; Pedersenet al. 1995). Tomato (Lycopersicon esculentum Mill., recently renamed Solanum lycopersicum L.) was used for this study because all 12 of its SCs are identifiable on the basis of relative lengths and arm ratios (Sherman and Stack 1992, 1995) and because tomato is a true diploid of agronomic importance (Rick 1991). Tomato SC 11 was chosen for hybridization because it is one of the shortest tomato chromosomes, and its corresponding linkage group contains 16 separate RFLP loci, several of which are associated with known genes (Tanksleyet al. 1992). Combined propidium iodide and 4′,6-diamidino-2-phenylindole (CPD) staining was employed to facilitate chromosome identification and aid in relating hybridization sites to chromosome structures. Our results suggest that FISH to SC spreads (SC-FISH) can be used to construct comprehensive maps of single-copy sequences on pachytene chromosomes.
MATERIALS AND METHODS
Screening the lambda tomato genomic library: Plasmids containing the tomato restriction fragment length polymorphism (RFLP) markers TG46, TG400, and TG523 were provided by S. D. Tanksley (Cornell University). TG523 flanks the jointless gene (Winget al. 1994), and TG400 is linked to the hairless gene (Tanksleyet al. 1992). To date, TG46 has not been closely associated with a particular gene or phenotype. While molecular and genetic evidence strongly suggests that TG46 and TG523 are single-copy markers, Southern blots of TG400 indicate that it may be present in more than one copy per haploid genome (S. D. Tanksley, personal communication; also see SolGenes web site, http://geneous.cit.cornell.edu/solgenes/aboutsolgenes.html/). Transformation of the competent E. coli strain DH5α, plasmid isolations, and verification of RFLP marker size by restriction analysis were performed using standard methods (Sambrooket al. 1989). The polymerase chain reaction was used to concomitantly amplify and label RFLP markers with digoxigenin (DIG; Rashtchian and Mackey 1992), and the resulting probes were used to screen a Clonetech (Palo Alto, CA) tomato lambda (λ) genomic library (cat. no. FL 1082d). Plating and filter preparation were performed as described in the Clonetech Lambda Library Protocol Handbook (PT1010-1). For each plate, an original and a duplicate filter were prepared. Saline sodium citrate (SSC) solutions of various strengths were produced by diluting a20× SSC stock solution (3 m NaCl, 0.5 m sodium citrate, pH 7.0) with deionized water. Filter hybridization and detection of positive clones were performed according to the Boehringer Mannheim (Indianapolis) DIG/Genius User’s Manual (see http://biochem.boehringer-mannheim.com/prod_inf/manuals/dig_man/dig_toc.htm/) with modifications as suggested by A. S. N. Reddy and I. S. Day (personal communication). Briefly, nylon filters (Boehringer Mannheim) were placed in standard hybridization buffer [5× SSC, 0.1% N-lauroylsarcosine, 0.02% sodium dodecyl sulfate (SDS), 1.0% w/v Boehringer Mannheim blocking reagent, 0.01 m maleic acid, 0.015 m NaCl, pH 7.5] for 1 hr at 55°, incubated overnight at 55° with gentle shaking in standard hybridization buffer containing 0.01 μg/ml heat-denatured DIG-labeled probe, washed twice in 2× SSC containing 0.1% w/v SDS (55°, 5 min each wash), washed twice in 0.5× SSC containing 0.1% SDS (55°, 15 min each wash), rinsed for 1 min in washing buffer (0.1 m maleic acid, 0.15 m NaCl, 0.3% Tween-20, pH 7.5), placed in blocking solution (0.1 m maleic acid, 0.15 m NaCl, 1% w/v blocking reagent) for 1 hr at 20° with gentle shaking, and incubated in a 1:2500 dilution of anti-DIG-alkaline phosphatase (750 units/ml, Boehringer Mannheim) in blocking buffer for 45 min (20° with gentle shaking). After incubation with the primary antibody, filters were washed twice in 500 ml washing buffer (15 min each wash), placed in detection buffer [100 mm Tris-(hydroxymethyl)-aminomethane, 100 mm NaCl, pH 9.5] for 2 min, and incubated in the dark without agitation in 200 ml of detection buffer containing 900 μl of nitrobluetetrazolium solution (75 mg/ml in 70% dimethylformamide, Boehringer Mannheim) and 700 μl of 5-bromo-4-chloro-3-indoyl phosphate toludinium salt solution (50 mg/ml in 100% dimethylformamide, Boehringer Mannheim). After ∼20 min, positive plaques (small blue “o-shaped” rings) began to appear on filters. Filters with a positive plaque in the same location on both the original and duplicate filters were thoroughly washed in deionized water, and corresponding plaques were removed from petri plates as described by Sambrook et al. (1989). Secondary and tertiary screens were performed as described above. For each of the three starting RFLP markers (TG46, TG400, and TG523), one corresponding lambda clone (designated λTG46, λTG400, and λTG523, respectively) was selected for use in FISH.
Probe preparation: The QIAGEN (Valencia, CA) Lambda Maxi Kit was used to isolate DNA from λTG46, λTG400, and λTG523. The DNA was digested with BamHI or EcoRI, and tomato insert DNA fragments were separated from λ arms by gel electrophoresis. Bands containing tomato insert DNA were excised from 1% w/v agarose gels, and the QIAGEN Qiaex II Kit was used to isolate DNA from agarose. The GIBCO BRL (Rockville, MD) BioNick Kit was used to label DNA with biotin. Each biotin-labeled probe was placed in its own 1.5-ml micro-centrifuge tube.
Tomato synaptonemal complex spreads: SC spreads were prepared as described by Peterson et al. (1996) with several modifications. Briefly, tomato (cv. Cherry) anthers were placed in 200 μl of sugar-salt medium [0.56 mm KH2PO4, 0.1 mm acid PIPES, 0.2% w/v potassium dextran sulfate, 1 mm CaCl2, 0.7 m mannitol, 1% w/v polyvinylpyrrolidone (Mr = 10,000), pH 5.1] containing 3 mg desalted cytohelicase (Sepracor, Marlborough, MA). The upper tips of the anthers were cut off, and anther bottoms were allowed to digest in the dark for 10 min. Drops of 45% acetic acid were placed on new microscope slides (Corning, Inc., Corning, NY), and slides were wiped dry with a Kimwipe. Slides then were glow-discharged according to Dubochet et al. (1971) except that air rather than amylamine vapors was introduced into the discharge chamber. Dissecting needles were used to squeeze microsporocytes from anthers, and microsporocytes were then allowed to digest for an additional 10–15 min. A 0.5-μl aliquot of the protoplast suspension was drawn into a siliconized micro-pipette and gently blown into a 10-μl droplet of hypotonic bursting medium (0.05% v/v Nonidet P-40 and 0.1% w/v bovine serum albumin) suspended from the end of a 200-μl pipette tip. The resulting droplet was immediately placed in the center of a glow-discharged (hydrophilic) microscope slide, and an additional 10 μl of hypotonic bursting medium was added. A hand-held nebulizer (Fullam, Latham, NY) was used to immediately give the slide 30 puffs of 4% aqueous formaldehyde (pH 8.5). The slide was air-dried, fixed in 4% aqueous formaldehyde for 10 min, rinsed twice without agitation in aqueous 0.01% Photoflo 200 (Kodak, Rochester, NY), rinsed four times (20 sec each rinse) in distilled water, and air-dried.
Fluorescence in situ hybridization: FISH was performed using a combination of the protocols of Rayburn and Gill (1985) and Anamthawat-Jónsson et al. (1996) with significant modifications. Slides with spread SCs were incubated in 2× SSC containing 100 μg/ml RNase A (Sigma, St. Louis) at 37° for 30 min, washed in two changes of 2× SSC (37°, 10 min each wash), rapidly dehydrated in a graded ethanol series, and air-dried. Chromosomal DNA was denatured by placing slides in 50 ml of 70% formamide in 2× SSC at 70° for 2.5 min. Slides were rapidly dehydrated in an ethanol series at -20° and air-dried. A cocktail containing 75 μl deionized formamide, 30 μl 50% w/v potassium dextran sulfate, 15 μl 20× SSC, 7.5 μl of 10 mg/ml sheared herring sperm DNA (GIBCO BRL), and 12 μlof25 μg/ml biotin-labeled probe(s) was heated at 97° for 10 min. Thirty microliters of the heat-denatured cocktail was placed on a slide, and a 22- × 50-mm coverglass was added. Slides were incubated in sealed humid chambers at 37° overnight. Coverglasses were gently washed from slides with 2× SSC at 37°. Slides were washed twice in 2× SSC, once in 25% formamide in 2× SSC, twice in 2× SSC containing 2% Tween-20 (all washes 37°, 10 min), and incubated for 60 min at 37° in blocking buffer (0.1 m maleic acid, 0.15 m NaCl, 1% w/v Boehringer Mannheim blocking reagent, pH 7.5). Slides were incubated for 1 hr at 37° in 1 μg/ml mouse anti-biotin (Boehringer Mannheim) in blocking buffer, washed three times in 2× SSC containing 2% Tween-20 (37°, 10 min each wash), incubated for 1 hr at 37° in 10 μg/ml biotinylated goat anti-mouse (Sigma) in blocking buffer, and washed three times in 2× SSC containing 2% Tween-20 (37°, 10 min each wash). Slides were then incubated in 40 μg/ml mouse anti-biotin fluorescein isothiocyanate (FITC; Sigma) or 4 μg/ml streptavidin-FITC conjugate in blocking buffer for 1 h (37°). All slides were washed once in 2× SSC containing 2% Tween-20 (37°, 10 min) and twice in 2× SSC at room temperature (10 min each wash).
Counterstaining and microscopy: Slides were incubated for 15 min in the dark in McIlvaine’s buffer (aqueous 17.7 mm citric acid, 164.7 mm Na2HPO4, pH 7.0) containing 0.5 μg/ml 4′,6-diamidino-2-phenylindole (DAPI), 1 μg/ml propidium iodide (PI), or 1 μg/ml PI and 0.5 μg/ml DAPI (i.e., CPD). Slides were washed for 30 sec in deionized water and allowed to air-dry for 5 min. Ten microliters of freshly prepared antifade medium (aqueous 50 mm Tris, 50% glycerol, 1 mg/ml phenylenediamine) was placed on each slide, and a 22- × 50-mm coverglass was added. Coverglasses were sealed onto slides with fingernail polish. Fluorescence and phase-contrast microscopy were performed using an Olympus Provis AX70 microscope equipped with an Olympus UM-51005 filter cube for simultaneous visualization of PI and FITC, a U-MNV filter cube for observation of DAPI and CPD staining, and a U-NG filter cube for visualizing PI alone. Photographs were taken using a PM-C35DX camera and Kodak Royal Gold 400 film.
CPD karyotype: Both freshly fixed SC spreads and SC spreads that had been used in FISH were CPD stained, and complete (or nearly complete) late pachytene sets were photographed using phase-contrast and fluorescence microscopy. Phase-contrast and corresponding CPD photographs (magnification of prints ×2258) were scanned at a resolution of 300 dpi into a computer. The chromosome-measuring program Micromeasure 3.01 (available at http://www.colostate.edu/Depts/Biology/Micromeasure/) was used to determine relative SC lengths and arm ratios from phase-contrast images. The SC karyotype of Sherman and Stack (1992, 1995) was used to identify individual-SCs. Adobe Photoshop 4.0 was used to digitally merge phase-contrast and corresponding CPD images. To describe the position of each CPD band, the distance along a CPD-banded SC from the center of the CPD band to the center of the kinetochore was measured and divided by the length of the entire SC arm. In this way, the position of each CPD band was expressed as a decimal fraction of the arm length from the centromere. These fractions were multiplied by the mean length of the SC arm (Sherman and Stack 1995) to give micrometer distances from bands to the centromere (Table 1). Likewise, the length of each CPD band was measured and divided by the length of its entire SC to determine the relative proportion of chromosome length contained within the band (Table 1).
FISH data analysis: For each late pachytene SC spread (i.e., spreads with visible kinetochores) showing a FISH signal, a phase-contrast and a PI/FITC photograph were taken. For most sets, a CPD image was also obtained. Phase-contrast and FITC/PI photographs were scanned and digitally merged to allow accurate measurement of distances between FISH signals and kinetochores. In all instances, SC 11 was identified on the basis of relative length and arm ratio according to the SC karyotype data of Sherman and Stack (1992, 1995). In spreads where FISH labeling was observed and a corresponding CPD image was available, the pattern of CPD staining was used to confirm the identity of SC 11 (see results). To determine the distance between an FITC focus and the kinetochore, the distance along the SC from the location of the focus to the center of the kinetochore was measured and divided by the length of the entire SC arm to express the focus location as a decimal fraction of the arm length from the centromere. Fraction distances were multiplied by the mean length of the SC arm to give micrometer distances from foci to the centromere. This is the same technique used by Sherman and Stack (1995) to describe the location of recombination nodules on tomato SCs. In cases where an FITC focus did not lie directly on the SC, a line was drawn from the center of the FITC focus perpendicular to the axis of the SC. The point of intersection between the line and the SC axis was marked as the location of that focus on the SC. In some instances, an FITC focus was located in the chromatin distal to the end of the SC. In these cases, a line was extended from the end of the SC, a perpendicular line was drawn from the focus through the extended line, and the point of intersection was marked as the location of the focus.
Karyotype of the CPD-banded tomato SC complement
Initially, all measurements were made from SC spreads labeled with only one of the three probes. However, after it was clearly established that there was a considerable physical distance between λTG523 and λTG46 and that there was no difficulty determining which signal belonged to which probe, measurements were made from these dual-labeled spreads as well. No measurements were made from spreads probed with any other combination of markers. Statistical analysis of the distribution of FITC foci for the three probes was performed using the computer program Instat 1.12a (GraphPad Software, San Diego).
Dot blot verification of probe identity: To rule out the unlikely possibility that the genomic probes used for FISH were inadvertently switched or cross-contaminated before in situ hybridization, each DIG-labeled RFLP marker was used to probe a separate nylon membrane on which a 20-μl drop (0.5 μg) of each of the biotin-labeled λ insert DNAs had been dotted. Filter hybridization conditions and colorimetric detection of probe hybridization were performed as described above.
RESULTS
DAPI, PI, and CPD staining of tomato SCs: When tomato SC spreads are stained with DAPI, pericentromeric heterochromatin, euchromatin, and telomeres are not readily differentiated. However, some bivalents possess a single, highly localized region that does not fluoresce, i.e., a DAPI-negative band (Figure 1A). DAPI-negative bands were observed throughout pachynema and diplonema. In late pachytene spreads, kinetochores stain with similar intensity as surrounding chromatin, which suggests that kinetochores are highly infiltrated with chromosomal DNA.
In PI-stained early pachytene SC spreads of tomato, heterochromatic regions often appear wider than euchromatic regions (Figure 1B). Additionally, heterochromatin stains a bit more intensely than euchromatin. In late pachytene spreads, PI-stained SCs have a rather homogenous appearance with no noticeable staining difference between heterochromatin and euchromatin. Kinetochores stain with an intensity similar to that of nearby chromatin, again indicating the presence of DNA in kinetochores.
When early pachytene tomato SC spreads are stained with CPD and examined using a wide UV filter, euchromatin and heterochromatin are easily differentiated with heterochromatin fluorescing white and euchromatin fluorescing blue (Figure 1C). The differential staining of heterochromatin and euchromatin seen after CPD treatment is more striking than differences produced by PI staining alone. By late pachynema, heterochromatin and euchromatin do not exhibit a predictable pattern of differential staining (Figure 1D).
Throughout pachynema, the NOR (which comprises most of the short arm of chromosome 2) is easily identified by its red fluorescence (Figure 1, C and D). As previously described, the NOR typically exhibits partial asynapsis (Sherman and Stack 1992; Petersonet al. 1996), and, as pachynema progresses, the NOR becomes more elongate (Figure 1D).
CPD staining results in a single red “CPD band” within the pericentromeric heterochromatin of 5 of tomato’s 12 SCs (Figure 1, C and D). The number and general location of CPD bands suggest that they are equivalent to the DAPI-negative bands seen in tomato SC spreads stained with DAPI only (Figure 1A). We used CPD staining coupled with the SC karyotype data of Sherman and Stack (1992), to construct a CPD karyotype for tomato. CPD bands are found on SCs 1, 3, 6, 8, and 11 (Figure 1, E–G and Table 1). The CPD bands on SCs 1 and 3 lie essentially at centromeric positions while the bands on SCs 6, 8, and 11 are located at subcentromeric positions in the short arms of their respective chromosomes (Figure 1G). An idiogram showing the distribution of CPD bands on the tomato SC complement is shown in Figure 1H.
—Staining of chromatin associated with tomato SC spreads. (A) Early pachytene SC set stained with DAPI only. Several SCs possess a region where DAPI staining is not visible, i.e., a DAPI-negative band (arrows). No other chromatin-based features are visible. (B) Early pachytene SC spread stained with PI only. Heterochromatic regions (e.g., arrowheads) appear wider and stain a bit more intensely than the distal euchromatic regions (e.g., arrows). (C) CPD staining of an early pachytene SC spread. Pericentromeric heterochromatin fluoresces white while distal euchromatin fluoresces blue. A red-fluorescing chromatin band (CPD band) is visible within the pericentromeric heterochromatin of 5 of tomato’s 12 SCs (arrows). These bands most likely correspond to the DAPI-negative bands seen in DAPI-stained tomato SC spreads. The NOR near the terminus of the short arm of SC 2 fluoresces red and usually exhibits partial or complete asynapsis (e.g., diamond-head arrows). (D) Late pachytene CPD-stained SC spread. Kinetochores are visible as thickenings along SCs (e.g., white boxes). On two CPD-banded SCs, the kinetochores are located directly over the CPD band (arrowheads). However, on the remaining three CPD-banded SCs a red band is visible outside of the region encompassed by the kinetochore (arrows). The elongate, asynapsed NOR fluoresces red (diamond-head arrows). (E) Phase-contrast image of a complete spread of late pachytene tomato SCs. A kinetochore is visible on each SC (black arrows). One SC is broken into two pieces (white arrow). Likewise, the NOR region has been lost from the end of SC 2 (white arrowhead). (F) CPD staining of the same SC spread shown in E. CPD bands are visible on five of the SCs (arrows). (G) Diagram showing the relative position of CPD bands and kinetochores. Blue lines represent SCs, yellow dots represent kinetochores, and red bars represent CPD bands. The identity (chromosome number) of each CPD-banded chromosome is shown next to an arrow pointing to its kinetochore. (H) Idiogram of the tomato SC complement showing the distribution of CPD bands. SCs 1-12 are shown in consecutive order from left to right. Euchromatin is blue, heterochromatin is white, kinetochores are orange, and CPD bands are red. The NOR in the short arm of SC 2, which appears red after CPD staining, is colored pink in this idiogram. A constricted area within the NOR marks the general region where SC formation does not occur. Frames that share a common bar: A and B; C and D; and E–G. Bars, 10 μm.
Fluorescence in situ hybridization: Probes containing single-/low-copy sequences were obtained by screening a tomato λ genomic library with the chromosome 11-associated RFLP markers TG46, TG400, and TG523 (Tanksleyet al. 1992). For each RFLP marker, a corresponding positive genomic clone was selected, i.e., λTG46, λTG400, and λTG523. Tomato insert DNA from each clone was isolated (lengths of tomato inserts: λTG46, 16,550 bp; λTG400, 10,200 bp; λTG523, 14,000 bp) and labeled with biotin using nick translation. Labeled insert DNA was hybridized to spreads of tomato SCs on microscope slides. Slides were treated with mouse anti-biotin followed by incubation in biotinylated goat anti-mouse. Sites of probe hybridization were detected using anti-biotin-FITC or streptavidin-FITC conjugates. SC 11 was identified on the basis of its relative length and arm ratio. Additionally, CPD banding analysis was used to confirm the identity of SC 11. The Micromeasure chromosome measuring program was used to determine the distance from the centromere to each FITC focus.
Each of the three genomic probes hybridized exclusively to euchromatin on the long arm of SC 11 (Figure 2). Little or no background FITC fluorescence was observed, and the efficiency of hybridization (i.e., the percentage of SC spreads where specific labeling of SC 11 could be seen) was high for all three probes (ca. 70%). On the basis of the relative locations of FITC foci, the three probes were positioned in relationship to the centromere and to one another (Table 2 and Figure 2, N–P). We refer to the diagram showing the position of each locus on SC 11 as an SC-FISH map (Figure 2O). One-way analysis of variance (ANOVA) indicates that the difference of the group means for the probes is highly significant (F = 146.9, P < 0.0001). Comparison of the data for any two individual probes also indicates that the loci are clearly separated (Bonferroni P value for any two probes is <0.001). λTG523 is the most distal, i.e., farthest from the centromere, of the loci with a near terminal location. The probes λTG46 and λTG400 are located more proximally on SC 11 with λTG400 being closer to the centromere than λTG46. This contradicts the molecular linkage map of chromosome 11 on which TG400 is located between TG523 and TG46 (Tanksleyet al. 1992; compare Figure 2N and 2O with 2P). To rule out the possibility that the probes λTG46 and λTG400 had been inadvertently switched before FISH, we performed a series of dot blot experiments. These experiments verified that each biotin-labeled FISH probe hybridized exclusively with the RFLP marker originally used to isolate it. It is unlikely that probes were switched at the time of FISH because data were obtained from several experiments, some experiments involved use of only one probe at a time, and probes used in all FISH experiments were removed from the same three tubes from which samples were taken for dot blot probe verification (see materials and methods).
SCs vs. “prefixed” pachytene chromosomes: In initial attempts to localize specific DNA sequences on meiotic chromosomes, we used pachytene chromosomes fixed before spreading in 1:3 acetic ethanol, i.e., “prefixed pachytene chromosomes,” as substrates for FISH rather than SC spreads. Although, when repetitive sequence probes were used, prefixed pachytene chromosomes were adequate substrates for FISH (e.g., Petersonet al. 1998), localization of single-copy probes on these chromosomes proved difficult because of high levels of background fluorescence and a scarcity of chromosome sets in which each bivalent could be distinguished, i.e., less than one completely interpretable set per slide (data not shown). When SC spreads were substituted for prefixed pachytene chromosomes, significant problems with background fluorescence were eliminated and chromosome identification was facilitated because each slide possessed 10 or more SC sets where all 12 tomato SCs were clearly identifiable (e.g., Figure 1, E–G).
DISCUSSION
CPD staining: If tomato SC spreads are stained with a combination of PI and DAPI, i.e., CPD stained, and illuminated with broad-band UV-visible light, structural features associated with differential chromatin condensation and/or DNA sequence can be visualized. These features either are not visible by DAPI or PI staining alone or are more readily differentiated by CPD staining. Although a variety of dye concentrations were tested, best results were obtained when chromosomes were stained with 1 μg/ml PI and 0.5 μg/ml DAPI (data not shown).
Euchromatin and heterochromatin can be differentiated by CPD staining in early pachytene SC spreads, but by late pachynema, the two types of chromatin are not readily distinguished. The explanation for this may be that the difference in relative condensation of heterochromatin and euchromatin is decreased during the transition from early to late pachynema. On the other hand, CPD bands and NORs remain differentially stained throughout pachynema (and probably through-out meiosis).
Likewise, a combination of PI and DAPI has been shown to differentially stain NORs in mitotic and meiotic metaphase chromosomes of cotton (Hansonet al. 1995; Jiet al. 1997), mitotic metaphase chromosomes of tomato (T. P. V. Hartman, personal communication), and SC spreads from maize and potato (our unpublished observations). Consequently, it seems likely that the NORs of many, if not all, plants can be visualized using CPD staining.
—Single-/low-copy FISH to tomato SC 11. (A) Phase-contrast image of a late pachytene SC 11. The kinetochore is clearly visible (arrowhead). (B) Hybridization of λTG523 to the long arm of the same SC. Probe hybridization is visible as yellow-green (FITC) foci (arrow). The chromosome appears red because of the PI counterstain. The kinetochore is marked by an arrowhead. (C) CPD image of the same SC. As predicted by the CPD karyotype, this chromosome possesses a red CPD band (arrow) in the short arm near the kinetochore (arrowhead). (D) Phase-contrast image of SC 11. The kinetochore is indicated by an arrowhead. (E) λTG523 hybridization on the SC shown in D. The signal is localized near the end of the chromosome (arrow). Note that there is little background FITC fluorescence, and there are no ectopic sites of hybridization. (F and G) Phase-contrast and PI/FITC image of a chromosome 11 showing the kinetochore (black arrowhead) and λTG400 hybridization (white arrow), respectively. (H and I) Phase-contrast and PI/FITC images of the same chromosome 11. In I, simultaneous detection of λTG523 (arrowhead) and λTG46 (arrow) is shown. (J–M) Centromere placement and probe positioning on an early pachytene tomato SC 11. (J) Phase-contrast image of an SC 11. No kinetochore is visible. (K) λTG523 hybridization (arrow). (L) CPD staining. Heterochromatin appears white and euchromatin blue. A distinct CPD band (arrowhead) is observed within the pericentromeric heterochromatin of the bivalent. (M) Diagram of chromatin features and hybridization sites superimposed on the phase-contrast image of the SC. In the diagram, euchromatin is blue, heterochromatin is green, the CPD band is red, and FITC foci are yellow. Although a kinetochore is not visibleon this SC, the consistent relationship between the location of CPD bands and kinetochores allows accurate placement of a “kinetochore” on the SC (black band; see discussion). (N–P) SC-FISH and molecular linkage maps of tomato chromosome 11. In all diagrams, λTG400, λTG46, and λTG523 are represented by the colors blue, red, and green, respectively. (N) Histogram showing the distribution of λTG46, λTG400, and λTG523 FITC foci along tomato SC 11. The number of observed FITC foci is shown on the ordinate axis while a diagram of an average SC 11 is shown on the abscissa. Each vertical bar represents the number of foci observed in a 0.1-μm interval on SC 11. On the SC 11 diagram, heterochromatin is represented by a gold rectangle encompassing the centromere (black circle), while euchromatin is represented by a thin, solid black line extending from both sides of the pericentromeric heterochromatin. A dotted line extends from the terminus of the long arm of SC 11 to include λTG523 foci that were found in chromatin extending beyond the end of the SC. (O) SC-FISH map of tomato SC 11. The mean position of each SC-FISH probe is shown by a vertical bar. All three probes are localized in euchromatin of the long arm. (P) Molecular linkage map of tomato chromosome 11 redrawn from Figure 1 of Tanksley et al. (1992). Regions between adjacent markers are enclosed by horizontal braces with map distances between markers given above the braces. Note that λTG46 and λTG400 are reversed in relation to their order on the SC-FISH map. Frames that share a common scale bar: (A–C) bar, 5 μm; (E–I) bar, 5 μm; (J–M) bar, 1 μm.
Locations of three probes on the long arm of tomato SC 11
Insight into the probable mechanisms by which CPD staining differentially stains NORs and CPD bands may be gained by considering the mechanisms by which PI and DAPI interact with DNA. PI intercalates between bases of single- or double-stranded nucleic acid molecules without regard to nitrogenous base composition, and consequently PI can be used as a quantitative nucleic acid dye (Heslop-Harrison and Schwarzacher 1996). DAPI is a double-stranded DNA-specific dye that interacts with DNA by at least two different mechanisms (see Kapuscinski 1995 for review). In regions where three or four AT base pairs are located in tandem, DAPI binds to the minor groove of the DNA, and this results in a highly fluorescent compound. DAPI also intercalates between bases, which results in a nonfluorescent compound. The former reaction is energetically favored in AT-rich regions while the latter is favored in GC-rich areas along a DNA molecule. As an intercalator, the binding strength of DAPI (in GC-rich areas) is roughly equivalent to that of PI. NORs, i.e., rDNA sequences, are GC-rich compared to most other DNA regions (Macgregor and Kezer 1971; Yasmineh and Yunis 1971; Ingleet al. 1975). As a result, in CPD-stained spreads, minor groove binding of DAPI to rDNA sequences would be limited. Presumably, DAPI and PI would compete for intercalation sites in the GC-rich rDNA, but only PI intercalation would result in fluorescence. In addition, the PI concentration in the CPD staining solution is double that of the DAPI concentration, and thus PI intercalation would be favored. Consequently, red (PI) fluorescence predominates in NORs. Similarly, DAPI-negative bands (seen in DAPI-stained SC spreads) and CPD bands probably represent other chromosomal regions rich in GC base pairs. The presence of GC-rich regions in heterochromatin (including GC-rich satellite sequences) has been reported for other species (see Sumner 1990 for review). Differential staining between heterochromatin and euchromatin observed in early pachynema likely reflects differences in both chromatin condensation and GC content.
CPD bands may be equivalent to certain C-bands. A C-band karyotype based on 1:3 acetic ethanol-fixed tomato (Moneymaker) pachytene chromosomes shows prominent centromeric/subcentromeric C-bands on chromosomes 1, 4, 6, 8, and 11 and staining of the NOR on chromosome 2 (M. S. Ramanna and L. P. Pijnacker, unpublished data). The relative location of major C-bands on chromosomes 1, 6, 8, and 11 corresponds exactly with locations of CPD bands on these chromosomes. However, the CPD- and C-band karyotypes differ in three ways: (1) There is no corresponding CPD band for the major C-band near the centromere of chromosome 4. (2) A C-band is found near the centromere of chromosome 3, but it is very small and probably does not account for the relatively large CPD-band we observe on this chromosome. And (3) there are several small C-bands for which corresponding CPD bands were not observed. These discrepancies may be caused by differences in the heterochromatin of the cultivars Cherry and Moneymaker. Also the mechanism of C-banding (Sumner 1990) differs from the proposed mechanism of CPD staining, so it is not surprising that some CPD bands coincide with C-bands while others do not (and vice versa).
SCs as substrates for single-copy in situ hybridization: Although FISH has been used to localize repetitive sequences on spreads of vertebrate SCs (Moens and Pearlman 1989, 1990; Henget al. 1994; Solari and Dresser 1995) and on SC spreads of two plant species (Hasenkampf 1991; Albini and Schwarzacher 1992), to our knowledge, there have been no reports of single-copy FISH to SCs of any species. Here we demonstrate that spreads of tomato SCs are well suited for the rigorous requirements of in situ detection of single-copy sequences. Because comparable SC spreads can be prepared for a variety of plants as well as animals and fungi, it is likely that this technique can be generally applied.
In our experience, background fluorescence is markedly reduced if SC spreads are used as FISH substrates instead of prefixed pachytene chromosomes. We suspect that this phenomenon reflects the time at which chromosomes are fixed. Treatment of cells with aceticethanol results in rapid protein crosslinking and fixation of chromosome structure (Sharma and Sharma 1980). Consequently, in prefixed preparations nucleoplasm may be fixed onto chromosomes before squashes/spreads are made. During FISH, probes and fluorochrome-labeled affinity reagents may become trapped within nucleoplasm surrounding the chromosomes to produce high background fluorescence. While digestion of chromosomes with a protease(s) before FISH can be used to remove at least some of the contaminating nucleoplasm from prefixed pachytene chromosomes (Jianget al. 1995), protease treatment also results in significant changes in chromosome structure (our experience). In contrast, SCs are spread before fixation. During the spreading process, nucleoplasm around SCs is dispersed and chromatin is decondensed, which allows better access of single-copy probes to chromosomal DNA.
Although we believe that SC spreads possess some features that make them better suited for single-copy FISH than prefixed pachytene chromosomes, it should be noted that other investigators have successfully localized single-copy sequences on prefixed pachytene chromosomes by in situ hybridization (Shenet al. 1987; Jianget al. 1995).
CPD banding and FISH: In all cases where single-copy FISH was detected and a CPD photograph was taken, a CPD band was observed in a position predicted by the CPD band karyotype of SC 11. Although SC 11 can be readily identified on the basis of relative length and arm ratio, FISH and CPD results each independently indicate that SC 11 was correctly identified in this study.
Nature of the tomato genome: Because it is possible that genomic clones isolated by screening a library with RFLP markers might contain repetitive elements in addition to single-copy DNA, genomic probes are often “prehybridized” with unlabeled repetitive DNA sequences before they are placed on chromosomes. Such chromosomal in situ suppression (CISS) hybridization effectively prevents highly repetitive sequences in probes from participating in hybridization with chromosomal DNA (Landegentet al. 1987; Lichter et al. 1988, 1990; Jiang et al. 1995, 1996; Zwicket al. 1997). However, in earlier studies we determined that only 23% of the tomato genome is composed of repetitive sequences and that the vast majority of these repetitive sequences are localized in pericentromeric heterochromatin (see Peterson et al. 1996, 1998). Consequently, because there was reason to believe that our genomic probes might not contain repetitive sequences, we tried FISH without CISS hybridization, and indeed, we found that each of the three genomic probes hybridized exclusively to a different highly localized region on SC 11 in the absence of suppression sequences. With a relatively small genome size and little repetitive DNA in its euchromatin, tomato should be particularly well suited for gene tagging and chromosome walking (Petersonet al. 1998).
Single- vs. low-copy number sequences: Molecular evidence indicates that TG523 and TG46 are located at single loci on tomato chromosome 11 (Tanksleyet al. 1992; SolGenes web site, http://geneous.cit.cornell.edu/solgenes/aboutsolgenes.html/). Our observation that λTG523 and λTG46 each bind exclusively to a single, highly localized region on SC 11 supports this premise. On the other hand, blotting evidence suggests that TG400 may occur at more than one locus in the tomato genome (S. D. Tanksley, personal communication; Sol-Genes web site). However, we observed only one highly localized site of λTG400 probe hybridization on tomato SC spreads, and this hybridization site is on the long arm of SC 11. If there is indeed more than one copy of TG400 in the tomato genome, this result might be explained if the copies of TG400 are so closely associated that they are not resolvable using our FISH method. Alternatively, it is possible that the λTG400 FISH clone hybridized to a TG400 site on SC 11 and did not recognize TG400 loci elsewhere.
Comparison of the SC-FISH and molecular genetic maps: While linkage maps show the relative order of genes along each chromosome, they provide little insight into the relationship between genes and chromosome structure. This is evident in a comparison of the SC-FISH map (Figure 2O) with the molecular linkage map of chromosome 11 (Figure 2P). On the molecular linkage map, TG523 is closer to the center of the map than it is to the terminus, but on the SC-FISH map, λTG523 is located near the end of the long arm of SC 11. Because TG523 is not the most distal locus on the molecular linkage map, it is likely that there are other genes/markers clustered within the more distal subtelomeric region of the chromosome. With respect to distances between markers, all three SC-FISH markers are contained within a chromosomal region comprising only 19% of the length of SC 11. However, the three RFLP markers span a region encompassing 35% of the molecular linkage map (Tanksleyet al. 1992).
On the SC-FISH map, the order of λTG46 and λTG400 is reversed in relation to their order on the molecular linkage map (Figure 2, N–P). This disagreement is not caused by an inability to resolve these markers cytologically, because (with the exception of a single FITC focus) the ranges of FITC foci for the markers do not overlap (Figure 2N). While somewhat unexpected, disagreement between the order of loci on tomato maps is not without precedent (e.g., Schumacheret al. 1995).
Relative order of TG46 and TG400: There are several possible explanations for the discrepancy between the relative positions of TG46 and TG400 on the SC-FISH and molecular genetic maps. For example, it is possible that (1) the λTG400 probe recognizes a locus on SC 11 different from the TG400 locus positioned on the linkage map (see above); (2) these loci have been incorrectly positioned on the molecular linkage map; or (3) the order of loci in cv. Cherry tomato differs from the gene order in the cultivar that was used to prepare the molecular linkage map (i.e., VF36-Tm2a; Tanksleyet al. 1992).
Tomato DNA per micrometer of SC in euchromatin and heterochromatin
The first of these possibilities seems unlikely because the λTG400 probe would have to consistently fail to hybridize to the “RFLP-mapped” copy of TG400 on chromosome 11, yet succeed in hybridizing to a second copy of TG400 on chromosome 11 overlooked during RFLP mapping. The homology between the two TG400 copies (even if limited to the 1200 bp of the RFLP marker itself) would probably be sufficient to produce some positive hybridization at both sites. Consequently, a less defined distribution of λTG400 FITC foci encompassing the range of λTG46 hybridization sites would be expected, but this is not observed (Figure 2N).
In reference to the second possibility, construction of linkage maps involves sexual crosses and sometimes difficult and/or complex scoring of recombination events that could lead to a mistake. However, because TG46, TG400, and TG523 have LOD scores of at least 3.0 (Tanksleyet al. 1992), it is unlikely that the markers have been positioned incorrectly on the molecular linkage map (see Tamarin 1999).
Perhaps the most likely explanation for the discrepancy between the SC-FISH and RFLP map is a difference in locus order between cv. Cherry and cv. VF36-Tm2a because of a small inversion in chromosome 11. In crosses between our Cherry tomato line and characterized translocation lines of cv. Early Fire Ball, no inversion loops were observed in pachytene SC spreads from the F1 progeny (unpublished observations made during research of Herickhoffet al. 1993). Consequently, if cv. Cherry contains an inversion in chromosome 11 relative to cv. Early Fire Ball, either the inversion is not large enough to produce an inversion loop in the heterozygote or cv. Cherry and cv. Early Fire Ball have the same inversion.
Base pair distances between SC-FISH markers: Data from previous investigations (Peterson et al. 1996, 1998) were used to calculate DNA densities (1C) in heterochromatic and euchromatic regions along tomato SCs (Table 3). This permitted an estimate of the number of base pairs between the centromere and each of the SC-FISH markers and between the SC-FISH markers themselves (Table 4). On the basis of the data in Table 4 and the average positions for the three probes shown in Figure 2O, we estimate that SC-FISH easily permits resolution of loci separated by as few as 1.7 million bp with every indication that much closer loci could be resolved. For comparison, Trask (1991) estimated that FISH probes can be mapped to human metaphase chromosomes at a resolution of ∼3 million bp. On the other hand, substrates other than condensed chromosomes can be used to improve resolution. For example, FISH has been used to resolve single-copy sequences separated by as little as 50 kb in the decondensed chromatin of interphase nuclei (Lawrenceet al. 1988; Trasket al. 1991). Additionally, Heng et al. (1992) resolved genes as close as 20 kb by FISH to free chromatin, and Fransz et al. (1996) resolved sequences separated by <1 kb using stretched, naked DNA as a FISH substrate. Unfortunately, with these latter techniques individual chromosomes and structural features such as heterochromatin, centromeres, and telomeres are not readily distinguishable.
Consolidation of the SC-FISH and recombination nodule maps: Because of the discrepancy between the SC-FISH and RFLP maps for tomato chromosome 11, a meaningful discussion of the relationship between molecular map distances and the number of base pairs between the three SC-FISH marker loci is not possible. However, the SC-FISH and recombination nodule maps for SC 11 (Sherman and Stack 1995) can be compared to determine map distances between SC-FISH markers. [Recombination nodules (RNs) are ellipsoidal structures associated with SCs during late pachynema, and there is a 1:1 correlation between RNs and crossover events (Carpenter 1975; Sherman and Stack 1995; Andersonet al. 1997).] Comparison of the maps indicates that the rate of crossing over per megabase pair (106 bp) is similar in the interval flanked by λTG400 and λTG46 and the interval flanked by λTG46 and λTG523 (Table 5). Thus, at least at the resolution of this analysis, there is no obvious recombination hot/cold spot in one of these intervals compared to the other (see Segalet al. 1992 for an example of a recombination hot spot in tomato euchromatin).
Use of CPD banding to position probes on early pachytene SCs: Because heterochromatin and euchromatin can be differentiated in early pachytene SCs, SC-FISH to early spreads may be useful in determining the chromatin background in which a gene/transgene is found. However, it is difficult to position hybridization sites on early pachytene SCs because they do not possess cytologically discernible centromeres/kinetochores at the light microscopic level (Stack and Anderson 1986). This problem can be overcome for chromosomes with a characterized CPD band(s) because the band can be used to position the centromere. An example of centromere placement in relation to a CPD band is shown in Figure 2, J–M. The SC 11 in these frames is 15.0 μm in length. The distance from the end of the short arm (i.e., the arm lacking a λTG523 FISH hybridization signal) to the CPD band is 5.8 μm. From Table 1, the relative distance between the CPD band and the kinetochore of SC 11 is 11.2% of the length of the short arm. Thus the distance from the terminus of the short arm to the kinetochore, i.e., the length of the short arm, is [5.8 μm ÷ (1.0 –0.112) =] 6.5 μm, and the long arm is (15.0 μm –6.5 μm =) 8.5 μm in length. Consequently, in Figure 2M the centromere (purple band) has been placed 6.5 μm from the end of the short arm. Once the centromere has been positioned, the locations of FISH hybridization signals can be expressed as fractions of the length of the chromosome arm to the centromere. For example, the two λTG523 FITC foci near the terminus of the long arm of SC 11 (Figure 2, K and M) are 7.8 and 7.9 μm from the centromere, respectively. Because the length of the long arm is 8.5 μm (see above), the two foci lie at points [100 × (7.8 μm ÷ 8.5 μm) =] 92.0% and [100 × (7.9 μm ÷ 8.5 μm) =] 93.0% of the long arm from the centromere.
Calculation of base pairs between different molecular and cytological markers
Map distances between markers determined by integrating the data from the recombination nodule map of tomato with the SC-FISH data
The potential of single-copy FISH to SCs: In this study we demonstrated that single-copy sequences can be reliably detected on SC spreads, discovered a discrepancy between our SC-FISH map and the molecular genetic map, estimated the distance between molecular markers in base pairs, and integrated the SC-FISH map with the preexisting RN map for tomato. Additionally, our data provide further evidence that there are few repeated sequences in tomato euchromatin. In broader terms, SC-FISH mapping has considerable potential as a means of relating genes to chromosome morphology (especially when coupled with differential chromatin-staining techniques such as CPD staining), studying the relationship between chromatin organization and gene expression, relating linkage distances to chromosome structure, and locating sites of transgene insertions. Because SC spreads have been prepared for many plants, animals, and a few fungi, it is probable that SC-FISH mapping can be used to investigate the genomes of a wide variety of organisms.
Acknowledgments
This research was funded in part by the United States Department of Agriculture grant 95-37300-1570 to S.M.S. and N.L.V. and Colorado Agricultural Experiment Station grant 1878-670 to S.M.S.
Footnotes
-
Communicating editor: W. F. Sheridan
- Received November 13, 1998.
- Accepted February 1, 1999.
- Copyright © 1999 by the Genetics Society of America
