|Home | About | Journals | Submit | Contact Us | Français|
The giant polytene chromosomes from Drosophila third instar larval salivary glands provide an important model system for studying the architectural changes in chromatin morphology associated with the process of transcription initiation and elongation. Especially, analysis of the heat shock response has proved useful in correlating chromatin structure remodeling with transcriptional activity. An important tool for such studies is the labeling of polytene chromosome squash preparations with antibodies to the enzymes, transcription factors, or histone modifications of interest. However, in any immunohistochemical experiment there will be advantages and disadvantages to different methods of fixation and sample preparation, the relative merits of which must be balanced. Here we provide detailed protocols for polytene chromosome squash preparation and discuss their relative pros and cons in terms of suitability for reliable antibody labeling and preservation of high resolution chromatin structure.
Drosophila has long been a favorite model system for studying the relationship between chromatin structure and transcription due to the cytological advantages provided by the giant salivary gland polytene chromosomes of third instar larvae. In this tissue the chromosomes undergo many rounds of replication in the absence of cell division giving rise to approximately 1000 copies. The DNA remains aligned after each replicative cycle resulting in greatly enlarged chromosomes. Using either phase contrast imaging or fluorescent microscopy of Hoechst-stained preparations, the more densely packed chromatin appears as bands whereas a more dispersed packing appears as interbands. The interband-specific localization of RNA polymerase II (Pol II) and associated transcription factors indicate that active genes tend to reside in interbands [1–5]. Further association of decondensed chromatin morphology with gene activity has been provided by studies of developmental- or stress-induced genes that show a “puffing” phenotype of the chromatin that correlates with high gene expression levels [6–9]. The increased decondensation, however, is not a direct consequence of gene expression as it can be uncoupled from transcription by chemical or promoter mutation methods [10, 11]. Indeed, dramatic remodeling of nucleosome architecture has been recently found to precede transcriptional activation after heat shock at the Hsp70 locus . Thus, polytene chromosomes provide a unique opportunity to examine both architectural changes in chromatin morphology as well as recruitment of specific enzymes and transcription factors involved in the process of transcription initiation and elongation.
Recent results have also underscored the complex choreography of different posttranslational histone modifications associated with regulation of transcription [reviewed in 13, 14]. Consequently, there has been a high level of interest in defining the modifications present at different genes and at different stages of the transcription process. An important tool for such studies is the labeling of polytene chromosomes with antibodies to the enzyme, transcription factor, or histone modification of interest. Here we provide various protocols for polytene chromosome squash preparation and discuss their relative pros and cons (summarized in Table 1) in terms of suitability for reliable antibody labeling and preservation of high resolution chromatin structure.
In any immunohistochemical experiment there will be advantages and disadvantages to different methods of fixation and sample preparation [15–18], the relative merits of which must be balanced. Thus, for any new antigen of interest, it is necessary to optimize the chosen fixative and fixation conditions. In the case of studies directed towards RNA polymerase II elongation control, the challenge is to identify conditions that suitably preserve the chromatin structure, allowing accessibility to antibodies without stripping away key proteins or blocking critical epitopes. Formaldehyde is a standard non-coagulative fixative choice that fixes tissues primarily by cross-linking, principally via lysine residues. Advantages include moderate (or efficient) penetration of tissues, partly due to the fairly slow kinetics of cross-linking, and provides stable covalent linkages . A disadvantage is that chromatin structure is not particularly well preserved in polytene squash preparations. Acetic acid is another popular non-coagulative fixative component known for its swelling affect on tissues  that in the case of polytene chromosome squashes helps to accommodate stretching of the chromatin in the interband regions . However acetic acid has a serious disadvantage in that it is prone to extract histones from the tissues [15, 21] and can “harden” the chromosomes, inhibiting their spreading. A solution to this problem was the inclusion of lactic acid in the fixative, which helped keep the glands "softer" and allowed for better unfolding of the chromosome arms [22, 23]. Still, many methods adopt a strategy of combining different fixatives with varying treatment times in an attempt to exploit the advantages afforded by a given agent while minimizing any inherent disadvantages .
Many studies of transcription and histone modifications rely on commercially available antibodies that often are poorly characterized and of variable quality. For these reasons Wang et al.  developed a “smush” technique based on a modified whole mount staining procedure for Drosophila third instar salivary gland nuclei that provides for a rapid and sensitive screening procedure. In this preparation nuclei from dissected salivary glands are gently compressed beneath a coverslip to flatten them before fixation in a standard paraformaldehyde/PBS solution of physiological pH that preserves the antigenicity of most epitopes. In a test case Cai et al.  recently used this technique to assess the suitability of different commercially available H3S10ph antibodies for immunohistochemistry on polytene chromosomes at interphase. The results showed that several of the tested antibodies from three different manufacturers were unreliable and that different lots of the same antibody had, in some cases, different properties . Although in the study of Cai et al.  the smush approach was used to identify reliable antibody lots, the ease and convenience of this approach would make it a powerful method to determine the suitability of different fixatives. In addition, for experiments in which it is important to preserve nuclear organization in its native state, for example to visualize the nuclear lamina, interchromosomal proteins, or the relative three-dimensional position of nuclear components, the “smush” technique would be the method of choice.
Although the resolution provided in smush preparations is sufficient to identify general patterns of chromatin proteins on chromosomes (Figure 1), the characteristic band/interband pattern of polytene chromosomes is not particularly well-resolved. This contrasts with conventional “squash” procedures where the strongly Hoechst-staining bands, consisting of more condensed chromatin regions, are distinct from the comparatively less-stained interband regions comprised of more “open” chromatin (Figure 2). The reproducibility of the polytene chromosome banding pattern allows chromosomal proteins such as polymerases and transcription factors to be localized to specific genomic regions. This presents a unique opportunity to study the molecular requirements for transcriptional regulation [1, 27, 28; reviewed in 29]. Analysis of the heat shock response has been a particularly good model system since it provides a well-characterized example of the induction of high levels of gene expression at specific “heat shock loci” with a concomitant downregulation of gene expression from other loci [9, 30, 31]. Antibody generated against RNA polymerase II that is phosphorylated at serine 2 in the C-terminal domain (Pol IIoser2) serves as a marker for active transcription [3, 32, 33]. Using this antibody the broad distribution of Pol IIoser2-labelled interbands visible under non-stressed conditions can be observed to change to a highly restricted pattern largely confined to the “heat shock puffs” after exposure to elevated temperatures (Figure 2). By immunostaining polytene chromosomes, recruitment and redistribution of a number of different proteins involved in the initiation/elongation process has been well characterized [5, 34–37]. However, such studies depend on suitable antibodies and fixation conditions that allow for the specific detection of these proteins.
The inclusion of acetic and lactic acids in the conventional squash fixation protocol facilitates both interband resolution and chromosomal arm spreading but unfortunately some epitopes do not survive this treatment. An example of such an epitope is H3S10ph (Figure 3A,B; also see ) which in contrast is preserved after formaldehyde fixation in the smush preparation (Figure 1). Since acid treatment also has the disadvantage that it quenches the inherent fluorescence of GFP-tagged proteins, DiMario et al.  recently developed a formaldehyde-based “acid-free squash technique” that allowed for direct visualization of GFP- fusion protein on polytene chromosomes without GFP-antibody labeling. In this case salivary glands were dissected directly in a dilute formaldehyde solution, soaked in 50% glycerol, and squashed in 50% glycerol in order to mimic the viscosity of the nucleus to help preserve chromosomal morphology during spreading. Preparation of the chromosomes using this technique allows the chromosomal arms to become extended during the squashing procedure and thus preserves the advantages of improved band/interband resolution of the squash technique. This is illustrated in Figure 4 where polytene chromosomes from a transgenic female larva expressing a JIL-1-GFP fusion protein  show robust GFP fluorescence that localizes to interband regions (Figure 4 top left pane) in the same pattern as has been previously observed using antibody-detection methods in female preparations [25, 39]. However, when squashes from the same transgenic line are prepared using a conventional squash protocol that includes acetic acid in the fixation buffer, all GFP fluorescence is lost (Figure 4 bottom left panel.) Furthermore, “acid-free” preparations also showed consistent localization of H3S10ph antibody labeling to interband regions with increased levels on the male X chromosome as expected (Figure 3C) in contrast to acid-treated preparations where labeling was severely attenuated and there was no detectable upregulation on the male X chromosome (Figure 3A; ).
Comparing Hoechst staining of acid and acid-free fixed preparations reveals another contrast between the two procedures: inclusion of acetic and lactic acids in the conventional squash technique enhances chromosomal spreading and interband resolution. Without the spreading advantages afforded by acid inclusion and since the cross-linking activity of formaldehyde tends to restrict chromosome arm spreading, it is essential to empirically optimize both formaldehyde concentrations and fixation times. As shown in Figure 3C, suitable conditions can be established such that comparably spread preparations can be obtained using the acid-free technique, but it is likely to take more squash attempts to obtain suitable samples than in the case of the conventional technique.
In some cases the resolution of the localization of chromatin proteins afforded by conventional squash protocols is still not sufficient to ascertain whether proteins co-localize and potentially interact. Recent developments utilizing different color variants of GFP-tagged proteins to image transcription dynamics on polytene chromosomes in live tissues along with potential applications such as FRET (fluorescence resonance energy transfer) to evaluate proximity of these proteins to each other  promises to provide a more detailed picture. However, higher resolution analysis extended to fixed preparations would be an extremely valuable tool in order to utilize many of the excellent antibodies that are already available. Recently a modified squash protocol featuring high pressure treatment that can yield chromosome spreads with resolution similar or equal to that of electron microscopy preparations has been developed . Furthermore, the use of a precision vise for the squashing step in this protocol not only allows increased pressure to be applied to the sample, but also significantly facilitates application of only vertical pressure during the squashing step, since any horizontal pressure at this stage is prone to shear the chromosome arms. This procedure significantly enhances reproducibility since squashing pressure is mechanically controlled and produces exceptionally flat chromosomes, accounting for the ultra-high resolution level. However, in some cases the thin-ness of the preparation also results in reduced antibody signal. In this case the investigator can adjust the pressure to be applied to optimize squash conditions for the desired purpose. As illustrated in Fig. 5 this approach provided further confirmation of the finding that JIL-1 histone H3 kinase is not associated with Pol II elongation , as the enhanced level of resolution afforded by this technique reveals minimal overlap between JIL-1 and Pol IIoser2 antibody labeling.
In a recent study analyzing H3S10ph staining of interphase chromosomes in smush preparation stainings, Cai et al.  found varying results for different commercial antibodies and even between different lots of the same commercial antibody . Thus it is essential to evaluate each individual antibody lot to determine its specificity and suitability for the intended uses. For example, after heat treatment, the high level of transcription of heat shock genes at 87A/C correlates with strong labeling by the actively transcribing Pol lIoser2 (Figure 2 and Figure 6). Anti-H3S10ph antibody from two different commercial sources (Epitomics and Cell Signaling Technology) show no labeling of the heat shock puffs (Figure 6A,B) . However, in contrast a third commercially available antibody (Upstate Biotechnology) showed strong labeling of the heat shock puffs (Figure 6C). In addition, strong labeling of heat shock puffs by the Upstate antibody was also observed in polytene chromosome squashes from JIL-1 null mutant larvae , which normally have undetectable levels of interphase histone H3S10 phosphorylation. Consequently, it is likely that the labeling of transcriptionally active heat shock puffs by the H3S10ph antibody (Upstate Biotechnology) is due to non-specific cross-reactivity, possibly with proteins involved in the heat shock response . These results underscore the importance of fully characterizing antibody specificity to ensure adequate performance.
“Antibody suitability” can also be strongly influenced by the experimental conditions selected and for that reason it is advisable to heed the manufacturer’s protocols for commercially available antibodies to optimize specificity for use for the preparation of interest. For example, the manufacturer’s instructions for immunoblot applications that accompany a commercially available antibody that detects Pol IIoser2 (Covance) specifically recommends against use of the popular blocking agent Blotto (5% nonfat dry milk) and instead specifies blocking with 5% BSA (bovine serum albumin). Figure 7 shows the results from immunoblots of larval salivary gland protein lysate fractionated by SDS-PAGE, electroblotted, and blocked using either 5% BSA or Blotto before probing with Covance anti-Pol IIoser2 antibody. Whereas antibody detection of the immunoblot that had been blocked in BSA as prescribed revealed a single band migrating at the appropriate molecular weight for Pol IIoser2, the immunoblot that had been blocked with Blotto showed, in addition to the Pol IIoser2 band, cross-reactivity with multiple different-sized bands.
Equally important, the specificity of the chosen secondary antibody should also be ascertained. For double labeling experiments the two separate primary antibodies used must be derived from different animal species or be of different isotypes in order to allow selective recognition by differentially-labeled secondary antibodies (e.g., one secondary antibody might be fluorescein-tagged while the other might be rhodamine-tagged). The secondary antibodies should be screened to ensure that they recognize only the relevant primary antibody and show no cross-reactivity to the other primary antibody. However, an additional concern is whether the secondary antibody might exhibit an unexpected cross-reactivity to antigen(s) present in the preparation. For example, we have encountered one case of a secondary antibody that shows a robust staining of transcriptionally active heat shock puffs on its own without addition of any primary antibody (Figure 8). Therefore control stainings using secondary antibody alone should always be performed to confirm absence of any such cross-reactivity. In some cases low levels of background reactivity can be eliminated by incubating the secondary antibody diluted 1:10 with fixed, devitellinized embryos in order to remove antibodies that might cross-react with Drosophila proteins. However, in most cases it is possible to screen different commercially available affinity purified antibodies to find one that does not show such cross-reactivity.
Drosophila are raised according to standard protocols . In order to obtain optimal polytene chromosomes for any of the techniques described below, uncrowded culturing conditions are essential (e.g., place around 20 egg-laying female flies in a bottle and change to new bottle each day). Select the fattest individuals from the first crop of climbing 3rd instar larvae while they are still wandering but just prior to pupation. We routinely culture at 21˚C but 18˚C will yield fatter chromosomes that may be more suitable for some purposes.
1˚ antibody (as determined experimentally)
2˚ antibody (for detection of the selected 1˚ antibody) [Note: optimal dilution should be empirically determined for each antibody lot. Select only affinity purified antibodies and test for cross-reactivity in the absence of primary antibodies. Filtering the diluted antibody through a syringe filter will remove fluorescent aggregates that may otherwise appear as speckles in the epifluorescent images.]
Hoechst 33258 (Molecular Probes)
Hoechst Solution (Hoechst 33258 0.2 µg/ml in PBS) (store in dark bottle at 4˚C) [Note: This concentration of Hoechst is less than recommended in most standard protocols, but in our experience results in superior labeling and resolution of band/interband regions.]
Nail polish for sealing coverslips [Note: use brightly colored nail polish instead of clear in order to observe whether the edges are fully sealed.]
We thank members of the laboratory for discussion, advice, and critical reading of the manuscript; Ms. V. Lephart for maintenance of fly stocks; and Mr. Laurence Woodruff for technical assistance. We thank Drs. M. Kuroda, R. Kelley, P. DiMario, and A.S. Belmont for advice and technical suggestions and for providing their protocols prior to publication. This work was supported by National Institutes for Health grant (GM62916) and National Science Foundation grant (MCB0817107) to KMJ.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.