Review Paper

Human Autoantibodies Reveal Titin as a Chromosomal Protein
Cristina Machado, Claudio E. Sunkel, and Deborah J. Andrew

     Titin is a large, flexible protein responsible for the elasticity of striated muscle and may also function as the scaffold upon which sarcomeres are assembled into myofibrils, the cylindrical elements that constitute the mass of muscle cells.  A PEVK domain and Ig/ FN3 repeats are thought to be the molecular basis for the spring-like nature of titin molecules.  A similar elastic nature has also been described in recent studies of chromosomes. This group cloned a gene in Drosophila, which they claim encodes the homologue of vertebrate titin, based on protein size, sequence similarity, developmental expression, and subcellular localization. The authors of this paper propose that a chromosomal form of titin could provide elasticity to chromosomes and resistance to chromosome breakage during mitosis.
     First, human cells and early Drosophila embryos were screened in order to identify human autoimmune sera that recognized nuclear components with cell-cycle dependent distribution.  Previous studies had shown that autoantibody targets of scleroderma, a multisystem connective tissue autoimmune disease, reacted with antigens localized to metaphase chromosomes and to the centrosome.  One human scleroderma serum was identified that stained chromosomes in both human cells and Drosophila early embryos.  Figure 1A is comprised of photographs that illustrate the chromosomal staining pattern that results when human epithelial cell nuclei (Hep-2 cells, on left ) and Drosophila 0-2h embryonic nuclei (on right) are stained with human scleroderma serum (green) and propidium iodide in order to detect DNA (red).  The propidium iodide stain was used as a positive control to identify all regions of the cells’ nuclei that contain DNA, and therefore chromosomes, for comparative purposes.  The merged image (yellow) indicates the region of overlap. In both sets of photographs, staining patterns show that the serum colocalizes with the total chromosomes in both types of cell nuclei.  During interphase, the chromatin is dispersed throughout the cells’ nuclei, and during prophase through telophase the chromosome staining condenses and becomes more localized.  This figure demonstrates that the human scleroderma serum stains mitotic chromosomes in both human cells and early embryo Drosophila cells.
   The human autoimmune scleroderma serum was then used to screen a Drosophila genomic expression library in order to identify the corresponding gene in Drosophila.  Five independent, overlapping clones were isolated that each contains multiple copies of a 71-amino acid repeat with an increased proportion of proline(P), valine (V), glutamic acid (E), and lysine (K) residues.  Figure 2C shows the amino acid sequence of the LG clone (the largest clone) rich with P, E, V, and K residues (63%).  This is similar to the PEVK-rich domain of the vertebrate titin, which provides muscle elasticity, which are 70% P,E,V, and K.
     Next, affinity-purified antibodies were generated against the LG protein (alpha-LG) and the KZ NH2-terminal peptide (alpha-KZ),  a protein expressed by one of the other isolated clones (KD clone.)  Figures 1B and 1C are photographs of HEp-2 cells (lefts panels) and Drosophila 0-2h embryos (right panels) stained with the alpha-LG (green) and propidium iodide (red) [top panels] and alpha-KZ (green) and propidium iodide (red) [bottom panels].  Merged images are located on the right (yellow.)  Both polyclonal antisera produce the similar chromosomal staining patterns  to those seen in human and embryonic Drosophila cells’ nuclei when stained with the human autoimmune serum.  In both figures 1B and 1C, condensed chromosome structure can be observed in metaphase and anaphase nuclei.  Together, these figures indicate that both antibodies and the human serum localize to the chromosomes in both types of cells.
     Figure 2a is a restriction map of the genomic region of the proposed D-titin gene.  Phage clones 1-9 were isolated by either using the LG genomic DNA expression clone as a probe or by using DNA flanking two nearby P-element insertion sites indicated by arrows, V(3)ET1 and V(3)ET2.  Three cDNAs mapping to discrete regions of the gene were isolated and designated KZ, NB, and JT according to the libraries in which they were found.  The largest cDNA is indicated as the JT cDNA, and its amino acid sequence is shown in figure 2C.
 The purpose of figure 2B is to illustrate the sequence homology between the chromosome-associated protein gene identified in Drosophila to those of vertebrate titins.  Figure 2B is a protein sequence alignment of corresponding open reading frames from two D-titin cDNAs (KZ and NB), chicken skeletal titin, and human cardiac titin.  An asterisk indicates identities among all three proteins and conserved residues between all three proteins are indicated by a period.  According to the authors, in the region of overlap, the ORF encoded by the KZ cDNA shows 28.6% identity/ 58.3% similarity to chicken skeletal titin, and 27.4% identity/ 56.8% similarity to human cardiac titin.  The ORF encoded by the NB cDNA shows 18.4%  identity/ 48.9% similarity to chicken skeletal titin, and 17.2% identity/ 47/7% similarity to human cardiac titin.  This figure suggests that the chromosome-associated protein isolated in Drosophila is homologous in sequence to those of vertebrate titins.
     Next, the in situ hybridization studies were performed in order to demonstrate whether the same or different genes encode the nuclear, chromosome-associated forms of titin and the muscle form of titin.  Pairs of embryos at the same developmental stage were analyzed by whole mount in situ hybridization to detect RNA levels on the left (dark blue) and by immunostaining to detect protein levels on the right (dark brown.)  Asterisks in Figure 2A indicate genomic fragments used to detect RNA accumulation in these studies, and the alpha-KZ was used in the immunostaining procedure.  The assumption was that if one of these probes or the antibody detected RNA or protein accumulation in the musculature of the developing embryo, then the same gene encodes for both the muscle and chromosome associated forms of titin.
 According to figure 3A/A’, Drosophila embryos show RNA accumulation and protein expression as early as the germ band extended stage (embryonic stage 10-11) in their musculature.  As can be seen from the staining pattern in figure B/B’ through H/H’, expression of RNA and protein continues to persist in some form throughout embryogenisis. D-titin is expressed in all the somatic, visceral, and pharyngeal musculature in some form throughout embryogenesis.  This evidence suggests that the same gene encodes for both the muscle and the chromosome associated form of titin.
     Then, the authors assessed whether or not the chromosome associated protein localized to specific regions within the sarcomere.  Antibodies directed against two different domains of the protein were used (alpha KZ and alpha LG) to stain Drosophila  adult thoracic myofibrils and larval gut muscle in Figure 4.  Panel 4a is a phage image of a myofibril from an adult thoracic muscle (top panel) stained with alpha-KZ (lower panel).  Arrows in the upper panel indicate Z disks, dense lines in the middle of each light band that separate on sarcomere from the next. Panel 4B is an adult thoracic myofibril double stained with alpha-KZ (green) and Texas red-phalloidin, which stains the filamentous actin of the I band. According to panels 4A and 4B, stained regions appear to correspond with the Z discs of each sarcomere.  Double staining with the alpha-KZ serum and either the human autoimmune scleroderma serum (Figure 4C) or the alpha-LG affinity purified antibodies (Figure 4D) also reveals Z-disc staining.  However, scleroderma serum also stained the M-line and along the length of the myofibril suggesting cross-reactivity to other antigens and/or the presence of additional, non-titin antibodies in the serum.  Figure 4E and 4F show the double staining of an adult thoracic myofibril with alpha-KZ (green) and  either MIR(red), an antibody that recognizes an epitope in the I-band near the I/A band junction, or anti-Zr5/Zr6, a polyclonal antiserum that was raised to the expressed alpha-actinin binding Z-repeat motifs Zr5/Zr6, respectively.  Figure 4G demonstrates that both alpha-KZ and alpha-LG antibodies stain the Z-discs of muscles from third instar larvae.
    An immunoblot was then performed to determine if D-TITIN is of an appropriate size to be the Drosophila homologue of the vertebrate muscle titin.  Figure 5 is a Western blot.  Proteins were run on a 2.5-7.5% denaturing gel and transferred to nitrocellulose.  Total protein extracts were taken from (a) Drosophila 8-24h embryos (after myogenesis), (b) Drosophila 0-2h embryos (several hours before myogenesis), and HeLa cells.  In figures 5a,b, and c, Lane 1 is stained with Coomassie blue and lanes 2-5 are immunoblotted with alpha-LG, LG preimmune serum, alpha KZ, and KZ preimmune serum, respectively.  Lane 6a is a shorter exposure of an immunoblot from 8-24h embryos incubated with the alpha-KZ serum, and it shows a "ladder-like array" of titin degradation products.  In all blots, lane 1 functions as a positive control being stained for total proteins, and lanes 3 and 5 function a negative controls revealing no cross-reactivity between polypeptides.  We would have like to have seen an actin control to ensure equal loading of protein between lanes.
     According to the authors, if D-TITIN is the homologue of vertebrate titin, then D-TITIN should be in the 2-4MD size range.  In Figure 5a lanes 2 and 4, discrete bands are detected in muscle cells with both the KZ and LG antibodies.  The authors claim that it is in the MD size range although it is difficult to determine due to the incomplete MW marker labels.  This is consistent with vertebrate titin.  Lanes 5b and 5c indicate that D-TITIN is also detected in non-muscle cells (0-2h embryos) and human epithelial cells (HeLa cells) as discrete bands of exactly the same size with both antibodies.  These blots provide further support that based on protein size, the Drosophila gene encodes the homologue of the vertebrate titin.
     Finally, condensed chromosomes were stained with antibodies to the vertebrate muscle titin.  In Figure 6, HEp-2 cells were double stained with antibodies to the vertebrate titin (green) and propidium iodide (red).  The merged image is located on the right (yellow).  Figure 6b is an illustration of the sarcomere structure, the titin domains, and the binding sites of the various antibodies used in the panels ofFigure 6a. Antibodies directed against the most NH2-terminal regions of the vertebrate titin, anti-Zr5/Zr and T12, did not detect titin on chromosomes (not shown); antibodies directed against the I-band regions of titin, N2A and 9D10 (not shown) showed weak chromosomal staining; the MIR serum (I/A-band junction) showed stronger chromosomal staining; and antibodies directed against A-band epitopes (BD6 and CE12) and the M-line epitope (A168) showed very strong staining of condensed chromosomes.  The two antibodies that did not recognize the titin on the HEp-2 chromosomes are directed against the NH2-terminal regions of titin that either map to the Z disk and bind to alpha-actinin.  This can be explained by the fact that the NH2-terminal region of the D-titin isoform encoded by the KZ cDNA does not contain the region homologous to the alpha-actinin-binding regions of vertebrate titin.  However, titin antibodies are proven to localize to chromosomes in Drosophila embryos, although the different forms of titin may vary in their NH2-termini.
     The data presented in this paper makes a strong case that the authors have indeed cloned the gene in Drosophila which corresponds to the vertebrate titin based on protein size, sequence similarity, developmental expression, and subcellular localization.  The data seems fairly solid, however, there are a few weak areas that should be pointed out.  Many of this group's experiments rely heavily on the use of two antibodies, alpha-KZ and alpha LG, and the use of the human autoimmune sera for staining.  The authors’ claim that these stains are specific based on the results in Figure 1, however, these experiments lack negative controls.  One negative control might be to stain whole cells versus just the cell nuclei. Most of the paper’s conclusions are based on the results of these stains.  These polyclonal antibodies and the antisera also may not be as specific to titin as the author’s claim.  Figure 4C is just one indication that the antisera and the antibodies may cross-react with other antigens/ and or the antisera may contain additional, non-titin antibodies.  However, despite some possible loopholes, the authors support their claim from various angles well.
     The authors then continue to state that "identification of titin as the chromosomal component provides a molecular basis for chromosome structure and elasticity."  This provides an open door for future experiments to test this hypothesis.  The next step would be experiments involving the genetic gain-of-function and loss-of-function studies using gene targeting by homologous recombination. One would expect that if the vertebrate titin gene were not being expressed, then cells would not be able to undergo mitosis properly.  Another step, that the paper alludes to, is to investigate whether or not vertebrate and muscle are splice variant of the same gene, or instead, are encoded by two closely related genes.  This question may be answer by deleting various portions of the gene and determining the presence or absence of protein product.