Lateral Elements Inside Synaptonemal Complex-Like Polycomplexes in ndt80 Mutants of Yeast Bind DNA

Corresponding author: Karin Schmekel, Stockholm University, Svante Arrhenius väg 16–18, SE-106 91 Stockholm, Sweden., karin.schmekel{at}molbio.su.se (E-mail)

Communicating editor: R. S. HAWLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The synaptonemal complex (SC) keeps the synapsed homologous chromosomes together during pachytene in meiotic prophase I. Structures that resemble stacks of SCs, polycomplexes, are sometimes found before or after pachytene. We have investigated ndt80 mutants of yeast, which arrest in pachytene. SCs appear normal in spread chromosome preparations, but are only occasionally found in intact nuclei examined in the electron microscope. Instead, large polycomplexes occur in almost every ndt80 mutant nucleus. Immunoelectron microscopy using DNA antibodies show strong preferential labeling to the lateral element parts of the polycomplexes. In situ hybridization using chromosome-specific probes confirms that the chromosomes in ndt80 mutants are paired and attached to the SCs. Our results suggest that polycomplexes can be involved in binding of chromosomes and possibly also in synapsis.

MEIOSIS, the process that organizes proper segregation of homologous chromosomes into haploid germ cells, depends strongly on chromosome pairing and homologous recombination. These two processes are essential for the homologous chromosomes to find each other, connect, and remain attached until they separate during the first meiotic division. The synaptonemal complex (SC) assembles between the homologous chromosomes during synapsis and keeps them paired until crossovers have established physical connections between them. The SC is a protein structure that consists of several components. Two lateral elements (LE), which are dense cores that run along the chromosomes, are directly attached to the chromosomes. The two LEs are connected with transverse filaments, while the central element is situated on the transverse filaments and between the LEs (SCHMEKEL and DANEHOLT 1995 ).

Polycomplexes are aggregates of SC-related material that are observed before SC formation, during regular SC formation, or most commonly, in postpachytene nuclei (reviewed in ZICKLER and KLECKNER 1999 ). Although the morphology of polycomplexes varies between species and between situations, polycomplexes generally appear to be stacks of aligned SCs with the LEs of the individual SCs fused. Polycomplexes appear frequently in many mutants of yeast that have deficiencies in chromosome synapsis, recombination, or both. The general view is that the LEs of the individual SC-like subunits in the polycomplex do not attach to the chromosomes. Although polycomplexes are often considered abnormal, they are seen in normal meiosis, as has been documented in several species, including yeast.

The Ndt80 protein is a meiosis-specific transcription factor in yeast that is expressed later than the earliest recombination genes and that regulates genes during the middle phase of sporulation (XU et al. 1995 ). Ndt80p is required for exit from pachytene and for full meiotic recombination.

In the present investigation we confirm that the SCs in spread preparations of ndt80 mutants appear similar to those in wild-type pachytene. Contrary to this impression is the fact that in situ prepared mutant cells essentially lack visible SCs as seen in electron microscopy (EM), but contain plenty of massive polycomplexes. The LE parts of the polycomplexes are preferentially stained by two independent DNA antibodies. Pairing of homologous chromosomes is confirmed in the mutant strain. We suggest that this is a situation in which SC entities inside polycomplexes bind chromosomes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains, growth, and spheroplast preparation:
The yeast (Saccharomyces cerevisiae) strains used in this study are as follows: NKY2310 (a/ ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/leu2::hisG ndt80::LEU2/ndt80::LEU2 his4X::LEU2-MluI-URA3/his4B::LEU2-MluI), NKY1551 [a/ ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/leu2::hisG arg4-Nsp/arg4-Bgl his4X::LEU2(Bam)-URA3/his4B::LEU2], and 1151(a/ ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/? COM1::Myc18::TRP1/COM1::Myc18::TRP1 trp1::hisG/trp1::hisG). The cells were cultured as described in SCHMEKEL 2000 . The cells were grown in sporulation medium at 30° for 5 hr (pachytene in the wild-type strains NKY1551 and 1151) or for 8 hr (the ndt80 mutant strain NKY2310).

EM preparation for serial sectioning:
For conventional transmission, EM preparations of spheroplasts were embedded as described in SCHMEKEL 2000 (p. 111). The sections were collected on copper slot grids coated with formvar and carbon and stained with 5% uranyl acetate in water for 1 hr and then with lead citrate for 1 hr.

Preparation for immuno-EM:
Embeddings for immuno-electron microscopy (immuno-EM) included preparation of nuclei according to SCHMEKEL 2000 (pp. 111–112), except that the cells were not DNase II treated. For embedding, the protocol in SCHMEKEL 2000 was used with the following modifications: The isolated nuclei were fixed in 2% paraformaldehyde in 0.1 M 2-N-morpholino ethane sulfonic acid (at pH 6.4) containing 0.5 M sorbitol, for 2 hr at room temperature (RT) and washed. For immunolabeling, the grids were incubated on drops of solutions: first for 10 min on 0.02% glycine in PBS to inactivate free aldehyde groups, then on 1% bovine serum albumin (BSA; Sigma A2153) in PBS for blocking, and 1 hr at RT. The grids were then incubated with the primary antibody in PBS containing 1% BSA. The sections were incubated with either of two monoclonal anti-double-stranded DNA antibodies (121-3 from BioGenex, San Ramon, CA, and MAB1293 from Chemicon, Temecula, CA). The cells were washed in PBS containing 1% BSA and then incubated with a gold-conjugated secondary antibody, gold/goat-anti-mouse (12 nm, Jackson ImmunoResearch Lab, West Grove, PA) in PBS containing 1% BSA. After washing, the grids were fixed in 4% glutaraldehyde and stained with uranyl acetate. The specimens were examined and photographed in a Zeiss EM 902.

Immunofluorescence and in situ hybridization:
The immunofluorescence experiments were performed with spread preparations on glass slides (described by LOIDL et al. 1998 ). The slides were incubated for 30 min with 3% BSA in PBS at RT and then incubated with a Zip1 antibody (2460; see description below) at a dilution of 1:100 in 1% BSA in PBS for 1 hr. The slides were washed in PBS and, as a last step before the secondary antibody incubation, in 1% BSA in PBS. Secondary antibody incubation was carried out with a FITC-conjugated swine-anti-rabbit serum (Dako, Glostrup, Denmark) for 30 min. DNA was detected by 4',6-diamidino-2-phenylindole (DAPI) staining. After washing in PBS, the slides were inspected in a fluorescence microscope (Zeiss Axioplan 2).

The fluorescence in situ hybridization (FISH) experiments were carried out according to WEINER and KLECKNER 1994 with minor modifications. The spreading was carried out according to LOIDL et al. 1998 . In brief, slides were treated with RNase A instead of both RNase A and RNase T. After denaturation the probe was applied to the slide and incubated for hybridization. Final washing was carried out in 1x SSC and 0.1x SSC. To enhance the hybridization signal, hybridization was repeated on one occasion. The two probes were produced from large pieces of chromosomes inserted into cosmids that were amplified in bacteria. One probe was from a subtelomeric region of chromosome III [c9171; American Type Culture Collection (ATCC) no. 70884], and the other was an interstitial part of chromosome VIII (c9315; ATCC no. 71216). The probe was purified using a Wizard plus midiprep DNA purification kit (Promega, Madison, WI) and digoxigenin labeled by nick translation according to standard procedures (Boehringer Mannheim, Indianapolis). The probe was detected using an antibody against digoxigenin (Boehringer Mannheim) as above and then by using a secondary antibody labeled with TRITC (Jackson ImmunoResearch Laboratories). The Zip1 protein was detected using the 2460 serum (see below) and a secondary antibody labeled with Cy5 (Jackson ImmunoResearch Laboratories). DNA was detected using a DNA dye, Yo-pro-1 (Molecular Probes, Leiden, The Netherlands). The images were recorded in a confocal microscope (Zeiss LSM 510).

The antiserum against Zip1 was produced by overexpression of a part of the ZIP1 gene on a plasmid (the same construct as that used for the production of the Zip1 antibody published in SYM et al. 1993 ). The glutathione S-transferase-tagged C-terminal part of the Zip1 protein was purified and injected into a rabbit. The serum (2460) stains SCs in spread preparations of meiotic nuclei and SCs in EM preparations and gives a single band in Western blots. The immunofluorescence and confocal microscopic images were processed using Adobe Photoshop, version 6.0.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The ndt80 mutant SCs appear as wild type in spreads:
As earlier documented, the ndt80 mutant yeast cells arrest in pachytene and appear more or less like wild type, as seen in spread preparations for light microscopy and EM (XU et al. 1995 ). We immunostained spreadings of ndt80 mutant and wild-type cells with a Zip1 antiserum for immunofluorescence microscopy (IF) and found that the SCs appeared similar in the two preparations. The ndt80 mutant cells had been cultivated in sporulation medium for 8 hr and had thus been arrested in pachytene for 3 hr. In almost all nuclei, SCs were clearly seen as individual entities (Fig 1, a and b), which colocalized with the labeling of DNA (Fig 1C and Fig D). In most nuclei of the ndt80 mutant we found bodies that were stained with the Zip1 antibody, which may be polycomplexes (Fig 1B, arrow). These Zip1 staining bodies were sometimes stained with DAPI, but usually that was not the case (Fig 2, a and b). Our overall impression is that the ndt80 cells appear similar to wild-type cells in their SC pattern.

View larger version (39K): In this window In a new window Download PPT slide   Figure 1. Fluorescence images of spread nuclei of wild-type (NKY1551; a and c) and ndt80 mutant (NKY2310; b and d). (a and b) Immunostaining with the Zip1 antiserum. (c and d) DAPI staining showing the DNA distribution. (b) In most ndt80 mutant nuclei, a majority of the SCs appear as separate entities; here all 16 SCs are distinguishable. The arrow points at a Zip1-staining body, probably a polycomplex. Note that this Zip1-staining body does not contain DNA. Bar, 2 µm.

View larger version (14K): In this window In a new window Download PPT slide   Figure 2. Fluorescence images of spread nuclei of the ndt80 mutant (NKY2310) immunostained with (a) Zip1-antiserum and (b) DAPI. Some nuclei show Zip1 staining that is similar to wild type (top cell of the three), while others show a body that stains heavily with the Zip1 antibody (the two nuclei in the bottom). In many cells, the body is negative for DNA staining by DAPI (the nucleus to the left, solid arrowhead), and in a few nuclei the body is stained by DAPI (the nucleus to the right, open arrowhead). Bar, 6 µm.

Intact cells of the ndt80 mutant lack visible SCs but contain large polycomplexes:
Spheroplasts that were cultivated the same way as those for the IF experiment were embedded for conventional EM. Ultrathin sections of wild-type nuclei in pachytene revealed individual, short pieces of SCs at pachytene with chromatin attached (Fig 3A). Surprisingly, the ndt80 mutant cells, on the other hand, showed massive polycomplexes in many nuclei (Fig 3B and Fig C). In the wild-type cells, polycomplexes containing two or three SCs were occasionally found in a single section (data not shown).

View larger version (79K): In this window In a new window Download PPT slide   Figure 3. Electron microscopic images of ultrathin sections of spheroplasts. (a) Short pieces of SCs in a wild-type cell. (b and c) Polycomplexes in a ndt80 mutant cell. Thin arrows point at the central element of SCs. Thick arrows point at spindle pole bodies. Bar, 300 nm.

ndt80 mutant cells were serially sectioned, documented in the EM, and reconstructed (the data are summarized in Table 1). To prevent biased selection, SC-containing nuclei, nuclei containing polycomplexes, and nuclei lacking any of the structures were chosen as starting sections. We found that 14 of the 16 nuclei contained massive polycomplexes. Four of the polycomplex-containing nuclei showed no free SCs, 8 nuclei had between one and five pieces of visible SCs in addition to the polycomplex, and 2 polycomplex-containing nuclei had seven SCs. Of the 2 nuclei that did not contain polycomplexes, 1 contained no SCs at all and the other contained two pieces of SC. Polycomplex-containing sections were counted and compared with the total number of sections through a nucleus to determine the fraction of sections that was occupied by polycomplexes. On average, 33% of the sections through a polycomplex-containing nucleus contained polycomplexes.

The number of nuclei that contain polycomplexes was counted in randomly selected single sections (N = 200). Eighteen percent of the nuclei contained polycomplexes, but since only 33% of the sections through a polycomplex-containing nucleus contain a piece of polycomplex, the actual number of nuclei that contain polycomplexes is 55%. A similar fraction of nuclei contained single pieces of SCs.

We have looked briefly at single sections of ndt80 mutant nuclei harvested after 5 and 6 hr in sporulation medium, but we have not done thorough investigations (i.e., serial sectioning) at these time points. At the earlier time points we found both SCs and polycomplexes, and the impression is that free SCs are more frequent and polycomplexes less frequent in those nuclei than in cells harvested after 8 hr.

In conclusion, very few single SCs are visible in the ndt80 mutant nuclei that have been arrested in pachytene for 3 hr, but most cells contain polycomplexes.

The SCs in the polycomplex are associated with DNA:
To further understand the chromosome pairing situation and the nature of these polycomplexes, we wanted to investigate the presence of DNA in the complexes. Ultrathin sections embedded for immunolabeling were incubated with either of two commercial antibodies (Chemicon and Biogenix) that bind to double-stranded DNA and then with a secondary antibody conjugated with 12-nm gold particles. Gold particles were located in the chromatin in general, in the nuclei and also in polycomplexes (Fig 4A). The gold particles in the polycomplexes were recorded and their precise location in the polycomplex was noted. Both antibodies showed strong preference for LE labeling (see Table 2). We noted that the outermost LE in a polycomplex appeared to have stronger labeling than the more centrally situated LEs and wanted to eliminate the possibility that the high LE labeling was due to the presence of more DNA in those LEs. We also wanted to eliminate the possibility that gold particles may represent the binding of the primary antibody to chromatin outside of the polycomplex. In the second counting, we therefore eliminated the gold particles in the outermost LEs and the outermost 30 nm of the polycomplexes in the transverse direction (see dotted square in Fig 4A). The limitation of 30 nm was chosen because this is the maximum distance from the antigen to the gold particle (BASCHONG and WRIGLEY 1990 ). Also under these conditions the labeling of the LEs was much stronger than that of the CE and the space between.

View larger version (82K): In this window In a new window Download PPT slide   Figure 4. Ultrathin sections of LR White-embedded nuclei of the ndt80 mutant strain NKY2310 (a) and strain 1151 (b) labeled with the anti-dsDNA antibody (Chemicon), marked by 12-nm gold particles. Arrows show gold particles (one single and a cluster) that are outside the polycomplex and that are not counted. Square enclosed by dotted lines shows the area that was counted in the more restrictive round of counting. Arrowheads show position of LEs. Bar, 100 nm.

Immuno-EM labeling of SCs in wild-type cells (the strain 1151) resulted in preferential labeling of chromatin and LEs, i.e., less in the central element area (Fig 4B). The SC morphology is less well preserved in the LR White-embedded cells than in the Agar 100 Resin-embedded cells, which are fixed with glutaraldehyde (Fig 3A), but the structure can be identified. Negative controls in which preparations of embedded nuclei were immunostained with the secondary antibody, in the absence of a first antibody, showed absence of labeling. In nuclei that had been DNase treated, the overall nuclear antibody staining, including the staining of the LEs, was very low even with high antibody concentration. One of the antibodies, the Biogenix, was totally negative after DNase treatment (data not shown). Thus, the immuno-EM investigation shows presence of DNA in the LE part of the SCs inside the polycomplexes.

The chromosomes are paired in ndt80 nuclei:
To see if the SCs in the ndt80 mutant cells keep the chromosomes paired/synapsed, we performed FISH experiments using chromosome-specific DNA probes. If the homologous chromosomes are paired, the signals should be fused into one dot in the confocal microscope (i.e., <0.7 µm apart; WEINER and KLECKNER 1994 ), and if the homologous chromosomes are not paired, two separate signals should appear. Hybridizations were done with the ATCC no. 70884 probe (subtelomeric, chromosome III) and with the ATCC no. 71216 probe (interstitial, chromosome VIII) on wild-type cells (NKY1551) and ndt80 cells (NKY2310). The ndt80 mutant cells showed a high level of pairing (the result is summarized in Table 3). Considering the fact that the wild-type cells are less synchronized than the arrested ndt80 mutant cells, it is not surprising that the mutants have a slightly higher degree of paired chromosomes. The SCs were well preserved in the FISH preparations and could be readily identified as outlined by the anti-Zip1 antibody staining (Fig 5). The localization of the probes was often distinctly seen at the location of the SCs, indicating that the SCs are attached to the chromosomes. Thus, the chromosome pairing in the ndt80 mutant appears like that in wild type.

View larger version (35K): In this window In a new window Download PPT slide   Figure 5. FISH experiment of spread preparations of ndt80 mutant cells combined with DNA labeling (green) and Zip1 labeling (light blue). (a) A nucleus with SCs labeled with the Zip1 antibody and probe 70884 labeling one end of chromosome III (red), showing the close location of the homologous sites on the two chromosomes. (b) The same nucleus with the DNA dye Yo-pro-1. Bar, 1 µm.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The phenotype of the ndt80 null mutants has previously been described, including the arrest in pachytene and the accumulation of SC-containing cells (XU et al. 1995 ). These observations have been done in hypotonically burst cells with the nuclei spread over a surface. In the same kind of preparations of mutant cells, harvested after 8 hr in sporulation medium, we also find wild-type-like SC formation. In addition, larger entities, possibly polycomplexes, are stained by the Zip1-antibody. We see that some, but not most, of the possible polycomplexes contain DNA.

Although the ndt80 mutants contain normal numbers of apparently normal SCs and although their homologous chromosomes are paired, we find almost no SCs in the EM in these cells. The cells were fixed and embedded according to a procedure that gives good SC morphology in wild-type cells. Instead we find massive polycomplexes in most mutant nuclei. We consider the most likely interpretation of our results to be that the SC entities in the polycomplexes seen in the EM are the same as the SCs seen in spreads. The discrepancy between the SC configuration in the two situations, solitary SCs in spreads vs. polycomplexes in intact cells, may be due to differences in specimen preparation methods. The spreading method is harsher than that of the EM embedding and may cause disruption of the polycomplexes, especially if the SCs are loosely connected.

One may think that the lack of visible SCs in the EM preparations of the ndt80 mutant is due to loss of morphological integrity of the SCs. However, the SC proteins cannot have been shed off the chromosomes, since at least one major SC component, the Zip1 protein, is still present in structures that appear as normal SCs in spreads. It further seems unlikely that the SC proteins have adopted an alternative organization inside the SC, because several examples of ndt80 mutant SCs are visible in EM and these look morphologically normal. Moreover, the normal central element-transverse filament-LE organization is seen in the polycomplexes that reside inside the mutant nuclei.

We find it likely that the SCs inside the polycomplexes bind chromosomes. The fact that the LEs inside the polycomplexes are preferentially stained with DNA antibodies strongly argues for the presence of chromatin inside the polycomplexes. Since the homologous chromosomes are closely situated in the arrested ndt80 nuclei, as shown by FISH chromosomal labeling on the SCs in spreads, we find it reasonable to believe that chromosomes inside the polycomplexes are synapsed. The DNA-containing Zip1 complexes that are seen in spread preparations may be remnants of the larger DNA-containing polycomplexes that we see in EM.

The polycomplexes in the ndt80 mutant probably originate from previously normal, solitary SCs that have later aggregated into polycomplexes. Preliminary observations that we have done at earlier time points indicate that the ratio of polycomplexes to solitary SCs gradually increases with time during the pachytene arrest.

Polycomplexes have usually been regarded either as aggregates of SC proteins under decay during stages when the SC disintegrates or as abnormal manifestations of meiotic mutations (reviewed in GOLDSTEIN 1987 ; ZICKLER and KLECKNER 1999 ). However, polycomplexes occur in wild-type cells, including yeast, during meiotic prophase I (ZICKLER and OLSON 1975 ; ZICKLER and KLECKNER 1999 ; our own observations). In intact wild-type yeast cells of the SK1 background, we occasionally observed polycomplexes containing two or three SC components in the EM.

Yeast cells go through meiotic prophase I rapidly. The meiotic stages between pachytene and metaphase (diplotene and diakinesis), in which the chromosomes gradually condense and organize into metaphase, are not very well defined in yeast as to chromosomal morphology. However, in a detailed EM study of wild-type (tetraploid) yeast meiosis, diplotene nuclei are described as lacking pachytene-like SCs and as having some SCs with thickened LEs and with polycomplexes in 30% of the cells (ZICKLER and OLSON 1975 ). We speculate that aggregation of SCs into chromosome-connected polycomplexes is part of a normal stage in the meiotic chromosome metabolism. The polycomplexes, as they appear in the ndt80 mutant, may be transient and less pronounced in wild-type cells. The formation of polycomplexes may be a transition from pachytene to later meiotic stages, perhaps involved in chromosome orientation.

	ACKNOWLEDGMENTS

We thank Dr. Nancy Kleckner and Dr. Franz Klein for yeast strains and Dr. Harry Scherthan for advice on the FISH method. We are grateful to Dr. Shirleen Roeder for giving us the ZIP1 construct from which we produced the antibody. Dr. Eva Bratt has generously given technical advice on confocal microscopy. We thank Drs. George Farrants and Christer Höög for constructive remarks on the manuscript. The project was supported by the Swedish Cancer Society, the Nilsson-Ehle Foundation, and the Borgström Foundation.

Manuscript received October 31, 2002; Accepted for publication November 14, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BASCHONG, W. and N. G. WRIGLEY, 1990  Small colloidal gold conjugated to Fab fragments or to immunoglobulin G as high-resolution labels for electron microscopy: a technical overview. J. Electron Microsc. Tech. 14:313-323.[Medline]

GOLDSTEIN, P., 1987  Multiple synaptonemal complexes (polycomplexes): origin, structure and function. Cell Biol. Int. Rep. 11:759-796.[Medline]

LOIDL, J., F. KLEIN, and J. ENGEBRECHT, 1998  Genetic and morphological approaches for the analysis of meiotic chromosomes in yeast. Methods Cell. Biol. 53:257-285.[Medline]

SCHMEKEL, K., 2000  Methods for immuno-electron microscopic and fine analysis of synaptonemal complexes and nodules in yeast. Chromosoma 109:110-116.[Medline]

SCHMEKEL, K. and B. DANEHOLT, 1995  The central region of the synaptonemal complex revealed in three dimensions. Trends Cell Biol. 5:239-242.[Medline]

SYM, M., J. A. ENGEBRECHT, and G. S. ROEDER, 1993  ZIP1 is a synaptonemal complex protein required for meiotic chromosome synapsis. Cell 72:365-378.[Medline]

WEINER, B. M. and N. KLECKNER, 1994  Chromosome pairing via multiple interstitial interactions before and during meiosis in yeast. Cell 77:977-991.[Medline]

XU, L., M. AJIMURA, R. PADMORE, C. KLEIN, and N. KLECKNER, 1995  NDT80, a meiosis-specific gene required for exit from pachytene in Saccharomyces cerevisiae. Mol. Cell. Biol. 15:6572-6581.[Abstract]

ZICKLER, D. and N. KLECKNER, 1999  Meiotic chromosomes: integrating structure and function. Annu. Rev. Genet. 33:603-754.[Medline]

ZICKLER, D. and L. W. OLSON, 1975  The synaptonemal complex and the spindle plaque during meiosis in yeast. Chromosoma 50:1-23.[Medline]

This article has been cited by other articles:

				F. Couteau and M. Zetka HTP-1 coordinates synaptonemal complex assembly with homolog alignment during meiosis in C. elegans Genes & Dev., November 15, 2005; 19(22): 2744 - 2756. [Abstract] [Full Text] [PDF]

				H. Bhuiyan and K. Schmekel Meiotic Chromosome Synapsis in Yeast Can Occur Without Spo11-Induced DNA Double-Strand Breaks Genetics, October 1, 2004; 168(2): 775 - 783. [Abstract] [Full Text] [PDF]