|Home | About | Journals | Submit | Contact Us | Français|
In this method we present biochemical procedures for rapid and efficient isolation of intermediate filament (IF) proteins from multiple mouse tissues. Isolated IFs can be used to study changes in post-translational modifications by mass spectrometry and other biochemical assays.
Intermediate filaments (IFs), together with actin filaments and microtubules, form the cytoskeleton - a critical structural element of every cell. Normal functioning IFs provide cells with mechanical and stress resilience, while a dysfunctional IF cytoskeleton compromises cellular health and has been associated with many human diseases. Post-translational modifications (PTMs) critically regulate IF dynamics in response to physiological changes and under stress conditions. Therefore, the ability to monitor changes in the PTM signature of IFs can contribute to a better functional understanding, and ultimately conditioning, of the IF system as a stress responder during cellular injury. However, the large number of IF proteins, which are encoded by over 70 individual genes and expressed in a tissue-dependent manner, is a major challenge in sorting out the relative importance of different PTMs. To that end, methods that enable monitoring of PTMs on IF proteins on an organism-wide level, rather than for isolated members of the family, can accelerate research progress in this area. Here we present biochemical methods for the isolation of the total, detergent-soluble, and detergent-resistant fraction of IF proteins from 9 different mouse tissues (brain, heart, lung, liver, small intestine, large intestine, pancreas, kidney, and spleen). We further demonstrate an optimized protocol for rapid isolation of IF proteins by using lysing matrix and automated homogenization of different mouse tissues. The automated protocol is useful for profiling IFs in experiments with high sample volume (such as in disease models involving multiple animals and experimental groups). The resulting samples can be utilized for various downstream analyses, including mass spectrometry-based PTM profiling. Utilizing these methods we provide new data to show that IF proteins in different mouse tissues (brain and liver) undergo parallel changes with respect to their expression levels and PTMs during aging.
IFs are a family of proteins that in humans are encoded by 73 genes and categorized into six major types: types I-IV are cytoplasmic (e.g. epithelial and hair keratins (K), myocyte desmin, neurofilaments, glial fibrillary acidic protein (GFAP), and others); type V are the nuclear lamins; and type VI are IFs in the eye lens1. In terms of their molecular organization, IF proteins have three common domains: a highly conserved coiled-coil “rod” domain, and globular “head” and “tail” domains. IF protein tetramers assemble to form short filament precursors, which are ultimately incorporated into mature filaments that shape dynamic cytoskeletal and nucleoskeletal structures involved in mechanical protection2, stress sensing3,4, regulation of transcription5 and growth, and other critical cellular functions1,6,7.
The functional importance of the IF system is highlighted by the existence of many human diseases caused by missense mutations in IF genes, including neuropathies, myopathies, skin fragility disorders, metabolic dysfunctions, and premature aging syndromes8. Some IF gene mutations do not cause, but predispose their carriers to disease progression, such as the simple epithelial keratins in liver disease9. The latter is due to the critical stress-protective functions of IFs in epithelia. IFs in general are among the most abundant cellular proteins under basal conditions, but are further strongly induced during various types of stress10. For example, recent studies evaluating proteome-wide changes in the nematode C. elegans demonstrated that multiple IFs are highly upregulated and prone to aggregation during organismal aging11,12. Since maintenance of a proper IF structure is essential for cellular resistance to various forms of stress10, IF aggregation may also contribute to the functional decline during aging. However, organismal-level studies examining multiple mammalian IF proteins across different tissues undergoing stress are lacking.
IFs are highly dynamic structures that adapt to meet cellular demands. Keratins, for example, undergo a biosynthesis-independent cycling between soluble (non-filamentous) and insoluble (filamentous) protein pool13. Under normal physiologic conditions approximately 5% the total K8/K18 pool can be extracted in detergent-free buffer, in comparison to approximately 20% that can be solubilized in the non-ionic detergent Nonidet P-40, which is biochemically comparable to Triton-X10014,15. During mitosis there is a notable increase in the solubility of simple-type epithelial K8 and K18,14 which is less apparent in epidermal keratins but more apparent in vimentin and other type III IF proteins 15,16. Solubility properties of IF proteins are tightly regulated by phosphorylation, a key post-translational modification (PTM) for filament rearrangement and solubility17–20. Most IFs undergo extensive regulation by a number of PTMs at conserved sites, resulting in functional changes.17
The purpose of this method is to introduce investigators who are new to the IF field to biochemical extraction and analytical methods for the study of IF proteins across multiple mouse tissues. Specifically, we focus on isolation of IF proteins using a high-salt extraction method and assessment of changes in PTMs via mass-spectrometry and by PTM-targeting antibodies. These methods build upon previously published procedures21 but include modifications for extracting different IF protein types to uncover common mechanism for regulation across the IF family. For example, K8 acetylation at a specific lysine residue regulates filament organization, while hyperacetylation promotes K8 insolubility and aggregate formation22. Recent global proteomic profiling studies have additionally revealed that most tissue-specific IF proteins are also targets for acetylation and that most IF acetylation sites are confined to the highly conserved rod domain (www.phosphosite.org). This highlights the need for methods suitable for global profiling of the IF system. We also introduce a rapid method of isolating IF proteins from multiple tissues using automated homogenization in optimized lysing matrix. The resulting preparations are suitable for downstream PTM analysis via mass spectrometry and other methods.
The protocol is approved and performed in accordance with the Institutional Animal Care and Use Committee (IACUC) at the University of North Carolina.
The traditional method25,26 of isolating the bulk of the intermediate filament protein fraction from epithelial tissue was modified here to include 9 different organs and a more rapid procedure for tissue lysis. While 3 manual homogenization steps are required for the traditional method, the modified procedure only has 1 manual homogenization step, which shortens the procedure by several hours, especially when processing more than 6 samples. Figure 1A shows a typical result of HSEs from 9 mouse tissues, while the table of expected proteins in the tissue and their molecular weight of each protein is shown in Figure 1B for comparison. Of note, the new automated method does not work well on pancreas and spleen; the traditional homogenization procedure works best for this tissue. This method should be useful for profiling IF proteins in a universal fashion across several tissues to assess organism-level responses to stress. As specific examples here, we show that IF proteins in liver (K8/K18) and brain (GFAP) from old mice are highly upregulated and undergo unique PTMs compared to the corresponding tissues in young mice.
Figure 2 shows typical results from several of the analyses described in this protocol. Panel A depicts expression analysis of the two major IF genes in the liver, keratin 8 (KRT8) and keratin 18 (KRT18), which encode the IF proteins K8 and K18, respectively. Epithelial keratins are strongly upregulated under various stress conditions. In the result shown, this occurs during aging, since KRT8 is significantly upregulated in the livers of 24 m old mice compared to 3 m old mice. The results at the protein level are more striking, as observed by the dramatic increase in K8 monomer as well as high molecular mass complexes in the old (24 m) livers. Coomassie-based protein stain here serves as a loading control to ensure equal loading of protein across samples. Note that with total tissue lysates it is easy to overload protein on a gel, as it is in this case. Loading a smaller volume or diluting the sample further in sodium dodecyl sulfate (SDS) buffer will alleviate this problem (especially if it appears viscous and difficult to pipet). Panel C depicts a typical result from a liver High Salt Extract obtained using the automated protocol, demonstrating the strong enrichment of keratins 8 and 18 on the gel. The red lines demarcate the area that was excised and submitted for mass spectrometry analysis. In panel D the results of the mass spec analysis of the samples in panel C shows that K8 and K18 in the old liver have multiple phosphorylation and acetylation sites that are not present in the young liver.
Figure 2 demonstrates that the methods used to extract IFs from epithelia can also be used on non-epithelial tissue. Furthermore, the results reveal a general pattern for aging-dependent upregulation of IF genes and proteins. The qPCR result in Figure 2A reveals a 5-fold induction of GFAP mRNA in the brains of 24 m old mice compared to 3 m old mice. Figure 2B reveals total proteins present in the Triton X-100 fraction and the IF-enriched HSE. Note the increase in the band intensity at 50 kDa marked by the arrow (corresponding to GFAP) in the old brain. Western blot analysis in Figure 2C further reveals the upregulation of GFAP and significant presence of GFAP monomer and high molecular mass complex in the Triton X-100 fraction (both marked by arrows). Western blot analysis of the same samples with a pan-acetyl lysine antibody shows that the antibody recognizes a band at ~50 kDa in the HSE of the old mouse brain and at ~250 kDa in the Triton X-100 fraction (Figure 2D). Immuno-enrichment of acetylated proteins from the Triton X-100 fractions reveals increased presence of GFAP protein in the lysate obtained from the old brain. Reducing and non-reducing conditions are shown to highlight GFAP monomer and high molecular mass complexes. Mass spectrometry analysis (similar to Figure 1) can be performed in this case to determine the site-specificity of the acetylated residues on GFAP in the old brain.
Methods that enable biochemical characterization of IF proteins can be useful to understand numerous pathophysiological phenomena in mammalian systems, since IF proteins are both markers and modulators of cellular and tissue stress29. The principle behind the current video method is based on the initial procedures developed in the 1970s and 1980s to isolate, separate and reconstitute IF proteins from cells and tissues, generally employing low and high salt solutions and Triton-X100 detergent25,30-35. For historical insight into the studies that supported the biochemical isolation of IF proteins please see recent reviews36,37. The current method is based on more recent protocols developed for the study of simple epithelial keratins26. The advantage of the video method is that it can serve as an initial step to enable investigators who are new to the IF field to effectively isolate IF proteins from most mammalian tissues. It represents a visual guide of related procedures that are widely used by investigators in the field to study IF protein regulation21,38.
This technique can be used to study the regulation and function of IFs in stress communication mechanisms between mammalian tissues39,40. It is becoming well-appreciated that stress in one tissue can affect the functioning of other tissues, for example under conditions of nutritional stress41, protein misfolding42, metabolic stress43, and other changes. Of note, most of the work currently being done in this area is in Drosophila and C. elegans models, while studies on widespread stress responses in mammalian systems are lacking. As the major regulator of cellular stress10, the IF system can unlock clues to organismal-level stress responses, which can have important disease implications. As such, the current protocol can be used to study global pathophysiological mechanisms in various mouse models of stress, injury and disease.
One limitation of the current technique is that the high salt extracts are considered “crude” IF preparations since they also contain other IF-associated proteins, such as plectin for example44. As shown previously, HSEs can be denatured in 8M urea buffer to obtain highly pure IF proteins that are capable of in vitro re-assembly in phosphate buffer44. After two such cycles of disassembly and re-assembly, the stoichiometry of IF proteins (isolated from HeLa cells) remained the same, whereas the urea treatment removed plectin44. During desmin purification, actin can be removed by solubilization of the HSE in acetic acid45. Whether additional purification steps should be included depends on the ultimate goal of the experiment. For example, if the goal is to characterize the assembly state and generate in vitro reconstituted IFs, the urea step is required to obtain highly pure IF proteins. On the other hand, if the goal is to identify context-dependent associations between IFs and other cellular proteins, then the “crude” preparations can be used. Interacting proteins can be identified by mass spectrometry using either in-gel digest for high abundance targets visible by gel stain, or in-solution digest for low abundance targets. Previous studies using immunoprecipitation of detergent-soluble IFs and HSEs containing IF proteins identified functionally important specific interaction between keratins 8/18 and heat shock protein 70 (Hsp70), which is potentiated after heat stress46, the adaptor protein 14-3-3 after phosphorylation47, and K8/K18 Raf-1 kinase under normal physiologic conditions48 among others. The current protocol can enable similar studies on other IF proteins, which are not commonly studied in biochemical settings, such as GFAP.
Use of the lysing matrix is a key modification from previous methods and should allow for rapid extraction of IFs from multiple tissues, with the exception of pancreas and spleen. The automation of the initial steps can be particularly useful for high-volume mouse experiments. For example, isolating IF proteins from 3 different tissues from 10 mice per condition (e.g. normal and some form of systemic stress or disease) will require processing of 60 individual tissues. Using the traditional manual method this would need to be done in batches and can take several days. However, using the automated method with lysing matrix, the procedure can be done in just a few hours. Critical steps of the procedure include: ensuring the sample stays cold throughout the procedure; incubation in high salt buffer for 1 h (step 4.5); using hot buffer and extensive vortexing to ensure complete resuspension of the pellet before analysis (step 4.8); and taking all precautions to avoid contamination of samples to be analyzed by mass spectrometry (step 7.2).
The major drawback to mass spectrometry-based identification of PTM sites is that it works very well for certain PTMs - phosphorylation and lysine acetylation being prime examples - but others, such as sumoylation are much more challenging49. Nevertheless, proteomics is a powerful tool for studying protein regulation in normal and diseased tissues. Agnetti et al. compiled a comprehensive up-to-date overview of cutting edge proteomic techniques currently in use to study various PTMs in the context of tissue-expressed proteins50. The purpose of the current method is to generate samples that can be applied for different types of analyses. For example, in vitro sumoylation can be performed on immune-precipitated IF proteins, and overall sumoylation can be assessed on isolated IFs in HSE preparations using sumo-specific antibodies18,21. Therefore the application of this method is not limited to generating samples only for mass spectrometry analysis but can be used to probe multiple IF properties using a variety of downstream techniques.
IF proteins are extensively regulated by numerous PTMs at conserved sites, resulting in altered solubility, filament assembly, and protein interactions17. Recent technological advances have unveiled the incredible magnitude, complexity, and disease significance of post-translational modifications (PTMs) on cellular proteins51–53. Improvement in detection methods for common PTMs, such as phosphorylation54 and acetylation51, have yielded vast catalogs of data, but understanding their biological meaning - cracking the “PTM code” - is a major remaining challenge55,56. One way to assess the functional roles and relative significance of each PTM is to determine how well it is conserved across multiple members of the IF protein family. For example, identification of a novel phospho-tyrosine site on K8 revealed that this site is contained within a QYE motif found in essentially all cytoplasmic IF proteins and critical for regulating IF protein solubility and filament dynamics57. Importantly, mutation of the conserved tyrosine residue on GFAP to a negatively-charged “phosphomimic” aspartic acid (Y242D) causes Alexander Disease58. The representative data shown here demonstrate how two different tissue-specific IF protein types (K8 and GFAP) undergo similar pattern of upregulation and increased acetylation during aging, which may be an effect of altered metabolism22,51. This example illustrates the utility of the method in understanding the function of IF proteins on a global physiological scale.
This work was supported by the NIH grants DK093776 [N.T.S.], DK102450 [N.T.S.], and P30 DK034987 [to UNC-Chapel Hill]. The authors thank Deekshita Ramanarayanan for assistance with qPCR and western blot experiments.
A complete version of this article that includes the video component is available at http://dx.doi.org/10.3791/55655.
The authors declare that they have no competing financial interests.
Rachel A. Battaglia, Department of Cell Biology and Physiology, School of Medicine, University of North Carolina – Chapel Hill.
Parijat Kabiraj, Department of Cell Biology and Physiology, School of Medicine, University of North Carolina – Chapel Hill.
Helen H. Willcockson, Department of Cell Biology and Physiology, School of Medicine, University of North Carolina – Chapel Hill.
Melinda Lian, Department of Cell Biology and Physiology, School of Medicine, University of North Carolina – Chapel Hill.
Natasha T. Snider, Department of Cell Biology and Physiology, School of Medicine, University of North Carolina – Chapel Hill.