Chicken Coagulation Factor XIIIA Is Produced by the Theca Externa and Stabilizes the Ovarian Follicular Wall*

Development of the follicle in egg-laying species such as the chicken is regulated by systemic factors as well as by the highly orchestrated interplay of differentially expressed genes within this organ. Differential mRNA display analysis of defined phases of follicle development resulted in the characterization of coagulation factor XIIIA. It is expressed and produced by cells of the theca externa in a highly regulated manner during distinct growth phases of the follicle. Transcripts for factor XIIIA are already detectable at the beginning of follicle development and peak at the end of phase 2. Protein levels, however, still increase during phase 3, peak shortly after ovulation, and persist until the postovulatory tissue is completely resorbed. Factor XIIIA is secreted as a monomer into the extracellular matrix of the theca externa and is not associated with factor XIIIB as is the case in plasma. Our data suggest that, due to its transglutaminase activity, factor XIIIA stabilizes the follicular wall by cross-linking matrix components. Thus, coagulation factor XIIIA might play a key role in coping with the massive mechanical stress exerted by the large amount of yolk accumulating during the rapid growth phase of the oocyte.

Reproduction in the mature hen depends on the coordinate differentiation and growth of oocytes. In the case of the domesticated chicken (Gallus gallus domesticus), fully developed oocytes are laid as eggs every 25 h. The oocytes develop in follicles, highly specialized structures of the ovary consisting of the oocyte proper in the center of the follicle, surrounded by concentric layers of cells, acellular material, and structures, including the perivitelline membrane (the equivalent of the zona pellucida in mammals), granulosa cells, a basement membrane, and the thecae interna and externa (1). Development of follicles can be divided into three major phases (for review see Ref. 2). Phase 1 is characterized by a very slow growth of follicles lasting for several months. At the end of this period, numerous follicles have reached a diameter of 2-3 mm; the oocytes within these follicles do not contain significant amounts of bona fide yolk (these follicles are referred to herein as small white follicles). During phase 2, some of these follicles develop further and reach a diameter of approximately 6 mm after 60 days. Due to yolk deposition into the oocyte, these follicles acquire a yellow appearance (here termed small yellow follicles). Finally, single follicles are selected from the pool of small yellow follicles every 25 h (synchronous with the ovulation cycle) and enter the rapid growth phase, which leads to mature follicles with a diameter of approximately 35 mm within 7 days. The fully developed oocyte is expelled from the follicle by ovulation and enters the oviduct, where egg formation starts. After ovulation, the remaining structure of the follicle, termed postovulatory follicle (POF), 1 consists of the granulosa cells, basement membrane, and theca layers. The POF collapses, but stays metabolically active for up to 6 days before involution and resorption (2). This developmental scheme establishes a hierarchy of follicles present at any given time in the ovary of a mature hen. Typically, the ovary contains five to eight prominent preovulatory follicles in the rapid 7-day growth phase. These follicles are numbered, from F1, the largest one, which will ovulate next, to the smallest distinguishable ones, usually F5 through F8.
Despite our detailed knowledge about vitellogenesis (for review see Refs. 3 and 4) and steroid production in the follicle (for review see Ref. 2), very little is known about the orchestration of these events at the molecular level. Especially, mechanisms responsible for establishing the follicular hierarchy are completely unknown. Additionally, another interesting point is the cellular and acellular architecture of the follicular wall. Although there is considerable knowledge about the morphology of this structure, very little is known about molecular and genetic aspects of cell differentiation and production of extracellular components during development of the follicle and its remodeling in preparation for and following ovulation.
We are approaching these questions from two angles. First, we have recently begun to study and characterize the production of the acellular structures within the follicles, perivitelline membrane, and the basement membrane. We found, e.g. that one of the zona pellucida proteins, ZPC, is synthesized by granulosa cells and deposited in polarized fashion into the space between granulosa cells and the oocyte (5). ZP1, another constituent of the perivitelline membrane, however, is synthesized in the liver, transported via the bloodstream to the ovary, and becomes deposited in the perivitelline membrane (6). Because follicle development takes place simultaneously, but nonsynchronously in many different follicles, we have to assume that the program guiding these complex mechanisms is not controlled by systemic factors, but rather within the follicle itself. We expect a highly orchestrated interplay of differen-tially expressed genes within different cells, dependent on growth phases of the follicle. Thus, as part of our second approach we have started to screen for such genes using the differential mRNA display methodology. Here we report results obtained in such a screen for genes expressed at various levels during different phases of follicle development. We show that the avian homologue of human coagulation factor XIIIA is expressed and secreted by theca cells and propose that it stabilizes the follicular wall by cross-linking matrix components during the last growth spurt of the oocyte, in preparation for ovulation of the follicle, for its subsequent transition into the postovulatory follicle and resorption.

EXPERIMENTAL PROCEDURES
Differential Display PCR-Total RNA from chicken follicles was prepared with TRI reagent (Molecular Research Center, Inc.). Poly(A) ϩ RNA was prepared with the Oligotex mRNA kit (Qiagen). 1 g of poly(A) ϩ RNA was mixed with 50 pmol of oligo-dT primer, incubated at 70°C for 10 min, and chilled on ice. First strand cDNA was synthesized in 20 l of buffer containing 2.5 M of each dNTP, 50 mM Tris-HCl (pH 8.3), 40 mM KCl, 6 mM MgCl 2 , 1 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, and 200 units of Superscript II reverse transcriptase (Life Technologies, Inc.) at 42°C for 50 min. The reaction mixture (20 l) for differential display (DD)-PCR contained 1 l of diluted cDNA (1:15), 1 M arbitrary primer (5Ј-TACAACGAGG-3Ј), 1 M anchored primer (5Ј-T 11 AC-3Ј), 2.5 M of each dNTP, 5 Ci of [␣-35 S]dATP, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 1.25 mM MgCl 2 , 2.5 units of Taq polymerase (Life Technologies, Inc.). After an incubation of 4 min at 94°C, the mixture was subjected to 40 low stringency cycles (94°C for 30 s, 40°C for 2 min, 72°C for 30 s) and a final extension at 72°C for 30 s. 5 l of the amplified cDNA fragments was separated on a 8% polyacrylamide 8 M urea gel. Bands were visualized by exposing the dried gel to x-ray film (Kodak X-Omat). Gel pieces containing the bands of interest were excised from the gel and rehydrated in 150 l of water by boiling for 30 min. Then the DNA was precipitated with ethanol and after dissolving in water used for reamplification as described for the original DD-PCR, except for the use of [␣-35 S]dATP. Reamplified DNA fragments were cloned into pCR2.1 (Invitrogen) and sequenced by the dideoxy chain termination method using the Sequenase Version 2.0 DNA sequencing kit (United States Biochemical). Sequence data obtained were subjected to homology search at nucleotide levels using the BLAST search of the National Institutes of Health.
Isolation of Chicken Factor XIIIA cDNA-DD-PCR clone 12 was used as a probe for screening a chicken follicle cDNA library (7) using standard conditions. Positive clones were subcloned into pBluescript II SK (Stratagene) and sequenced. To obtain the 5Ј-end of the coding sequence, 5Ј-rapid amplification of cDNA ends (RACE) was employed using the Marathon cDNA amplification kit (CLONTECH). The primers used were 5Ј-TCCTTCTCGTCATCCAGGTACACAGC-3Ј and 5Ј-T-CAGCACATCACCTATGAGGATCGG-3Ј.
Expression of Chicken FXIIIA-The full-length cDNA of chicken FXIIIA was cloned into the XbaI and SmaI sites of pCIneo (Promega) and expressed in the human embryonic kidney cell line 293. Transfection of the cells was performed using Lipofectin reagent (Life Technologies, Inc.). Stable transformants were selected by the addition of 500 mg/liter G418 to the medium (Dulbecco's modified Eagle's medium (Life Technologies, Inc.), 10% fetal calf serum, 584 mg/liter glutamine). Cells expressing factor XIIIA were washed three times with 2 ml each of phosphate-buffered saline (PBS), scraped from the dishes with a rubber policeman, and centrifuged at 2000 ϫ g for 5 min, and the cell pellet was solubilized by addition of 150 l of buffer/dish (200 mM Tris maleate, pH 6.5, 2 mM CaCl 2 , 0.5 mM phenylmethylsulfonyl fluoride, 2.5 M leupeptin, and 1.4% Triton X-100). The cell extracts were centrifuged at 300,000 ϫ g for 40 min at 4°C and the pellet discarded.
Purification of Glutathione S-Transferase Fusion Proteins and Antibody Production-cDNA fragments encoding amino acid residues 1-45 (corresponding to the activation peptide) and amino acid residues 674 -735 (corresponding to the carboxyl terminus) of chicken factor XIIIA were cloned into pGEX-5X-1 (Amersham Pharmacia Biotech). Isolation of the glutathione S-transferase fusion proteins in DH5␣ cells (8) and production of the respective polyclonal antibodies (9) were carried out as described in the indicated references. The antisera were designated Ac16 (against the activation peptide) and A3 (against the carboxyl terminus).
Immunoprecipitation and Western Blotting-For immunoprecipitation, 1 ml of conditioned medium (from 293 cells expressing FXIIIA and mock-transfected cells) was incubated for 24 h at 4°C with 5 l of antiserum (Ac16) or preimmune serum and with 4 g of Protein A-Sepharose (Amersham Pharmacia Biotech) in 10 mM Tris/HCl, 150 mM NaCl, 2 mM EDTA, pH 7.4. Then the Sepharose beads were washed five times with ice-cold PBS. Proteins were eluted from the beads with Laemmli sample buffer containing ␤-mercaptoethanol (5%, v/v) and separated by SDS-polyacrylamide gel electrophoresis. Presence of chicken factor XIIIA in the immunoprecipitates was tested by Western blotting as described (10).
Immunofluorescence-Follicles (6 mm) from an adult hen were dissected in ice-cold PBS, embedded in freezing agent (Microm, Austria) and immediately frozen. Cryostat sections of 12-m thickness were prepared, transferred to Superfrost-Plus slides (Menzel) and stored at Ϫ70°C. Slides were washed in PBS, incubated for 24 h with the respective primary antibody (Ac16; anti-␣-collagen I, Sigma; anti-chickenfibronectin, Chemicon Int.; anti-human-factor XIIIA, Calbiochem), washed, and incubated with a fluorescein isothiocyanate-conjugated anti-rabbit IgG (Sigma). After a final wash, microscopy was performed on a Zeiss microscope (Axiovert 135).
In Situ FXIIIA Activity Assay-Transglutaminase activity was detected as described previously (11). Briefly, slides prepared as described above were washed with PBS, preincubated in 100 mM Tris-HCl, pH 8, 1% bovine serum albumin, 10 mM CaCl 2 or 10 mM EDTA for 1 h at room temperature, and incubated with fluorescein-labeled cadaverine (Molecular Probes) in blocking buffer for 1 h at room temperature. After washing, microscopy was performed on a Zeiss microscope (Axiovert 135).
FIG. 1. Differential display of mRNAs from different developmental stages of chicken follicles. Poly(A) ϩ RNA from three different stages of follicle development were subjected to differential mRNA display analysis as described under "Experimental Procedures." A partial autoradiogram showing two selected mRNAs (#12, #13) is presented in A. B and C demonstrate the expression pattern of #12 and #13 as assessed by Northern blotting. Relative intensities of the bands normalized to the amount of 18 S rRNA are shown in the corresponding diagrams underneath the Northern blots. sw, small white follicles; sy, small yellow follicles; lw, large white follicles.

RESULTS
In a first step to identify genes involved in the progression of follicle development, mRNAs isolated from three distinct pools of differently staged follicles (small white, Ͻ2 mm; large white, 2-4 mm; small yellow, 4 -8 mm), covering phases 1 and 2 of follicle development, were subjected to differential mRNA display analysis (12). We have isolated several fragments derived from genes that appeared up-or down-regulated as judged from the intensities on the gel, e.g. fragments 12 and 13 displayed in Fig. 1A. However, only six of these fragments (derived from mRNAs displayed using the given primer combination) exhibited differential expression, as evaluated by Northern blotting (Fig. 1, B and C). To obtain sequence information from the coding regions corresponding to the selected fragments, these fragments were used to screen a chicken follicle cDNA library (7). BLAST searches against available data bases demonstrated that only one of the clones (number 12) was identical with a known chicken gene; it codes for a chicken chondrocyte transglutaminase (13,14). Thus, we focused our attention on clone 12, and the sequence of the corresponding full-length transcript with a length of 4.5 kilobase pairs was obtained by 5Ј-RACE. The sequence defined an open reading frame of 2208 base pairs (GenBank accession number AJ278103). Sequence alignments (Fig. 2) clearly identified the obtained cDNA as the chicken homologue of human factor XIIIA (15). The ATG serving as the translational start site was assigned based on the following considerations: First, it is located in a sequence context fulfilling the rules for a translation initiation site (16), and second, the corresponding chicken protein starts with the same sequence motif as human factor XIIIA (Fig. 2). The overall identity at the amino acid level between chicken and human factor XIIIA is 65%, displaying 100% identity around the active site cysteine, which catalyzes the acyl transfer reaction common to all transglutaminases (for review see Ref. 17).
Mammalian plasma factor XIIIA is part of a large tetrameric complex consisting of two molecules of XIIIA and two molecules of XIIIB. The cellular form of factor XIIIA, however, is a homodimer consisting of two A subunits only. An interesting functional feature of human factor XIIIA is a thrombin cleavage site between Arg-37 and Glu-38 (see Fig. 2). Cleavage of FXIIIA by thrombin at this site leads to liberation of the aminoterminal so-called activation peptide and subsequently to the dissociation of the A and B subunits. Thus, in the absence of the inhibitory B subunits, the thrombin-modified A subunits as- sume a Ca 2ϩ -dependent active conformation catalyzing the cross-linking of fibrinogen to insoluble fibrin.
To characterize the corresponding chicken protein further, we produced two antibodies directed against the putative activation peptide at the amino terminus (Ac16), and against the extreme carboxyl terminus of chicken XIIIA (A3), respectively. Using these antibodies in combination with a commercially available antibody against human XIIIA, we first characterized the plasma form of chicken XIIIA (Fig. 3A). Under non-reducing conditions, Ac16 recognizes all three forms of chicken factor XIIIA in plasma, i.e. the heterotetramer (2A2B), the homodimer (2A), and the monomer (A), respectively (lane 1). The occurrence of both the homodimer (2A) and the unclipped monomer (A) in plasma is due to the inherent instability of the tetrameric complex during electrophoresis. Under the same conditions, A3 and the commercial anti-human-XIIIA antibody recognize the heterotetramer in plasma (lanes 3 and 5). Independently of the enzyme's activation status, A3 reacts with the dimer as demonstrated by comparing plasma versus serum samples electrophoretically separated under non-reducing conditions (lanes 3 and 4). A3 only weakly recognizes the monomer of chicken XIIIA under the conditions tested (the monomer becomes visible only after prolonged exposure, lane 3). Removal of the activation peptide and dissociation of the A and B subunits was demonstrated by the absence in serum of any detectable heterotetramer using A3 and anti-human XIIIA (lanes 4  and 6). In addition, under reducing conditions, all of the XIIIA present in plasma (i.e. not activated) migrates as monomer and is strongly recognized by Ac16 (lane 7). Upon activation, in serum, hardly any reactive band is visible under reducing conditions, indicating that the activation peptide is removed (lane 8). Upon prolonged exposure of the blot, however, a faint band with a size similar to that of the monomer becomes apparent; most likely it represents a small residual amount of non-activated XIIIA (also visible as dimer under non-reducing conditions in lane 2). Thus, activation of chicken factor XIII, as it occurs during blood clotting, obviously follows a similar process as described for mammals (17).
Having characterized the circulating forms of chicken factor XIIIA, we studied the protein produced by chicken follicles. Under both reducing and non-reducing conditions, the only band detected by Ac16 is the monomer, indicating that XIIIA is not associated with the B subunit and that follicular XIIIA does not form a dimer as it does in the circulation (Fig. 3B, lane 1). This observation is substantiated by using A3, which reacts strongly with the dimer (2A) and the heterodimer of XIIIA (2A2B), but only weakly with the monomer (A), as demonstrated on plasma and serum samples. In the follicle this antibody visualizes only the monomer under the conditions applied (Fig. 3B, lanes 3 and 4).
Human factor XIIIA does not contain a hydrophobic leader sequence and thus belongs to a group of proteins which are secreted from cells by an "unusual secretory pathway" (for review see Ref. 18). Chicken factor XIIIA also does not contain a hydrophobic leader sequence (Fig. 2). To test whether chicken factor XIIIA can be secreted from cells, we expressed the fulllength cDNA in 293 cells and tested for the expressed protein in cell extracts and the medium. As seen in Fig. 4A, transfected 293 cells express large amounts of monomeric chicken factor XIIIA (lane 2). Again, antibody A3 does not detect a band corresponding to the homodimer seen under the same conditions in the plasma (Fig. 4, lanes 4 and 6). Furthermore, significant amounts of the non-activated monomeric protein are indeed secreted into the medium, as demonstrated by immunoprecipitation experiments using Ac16 (Fig. 4B, lanes 1-5).
To evaluate the expression pattern of XIIIA in chicken tissues, we examined all major organs (follicle, eye, intestine, liver, kidney, muscle, brain, heart, lung, skin, adrenal, and spleen) and relevant cell lines by Northern blot analysis (data not shown). The only site in female chicken expressing detectable transcripts for factor XIIIA is the follicle and, of the cell lines tested, HD11 cells, a chicken macrophage cell line (19).
From these experiments we conclude that in the chicken circulating factor XIIIA is highly homologous to human factor XIIIA and that it is associated into a tetrameric complex constituting plasma factor XIII in a similar way as described for the mammalian system. Expression analysis did not reveal where the circulating protein is produced. This situation is reminiscent of that in humans where the question about the site of synthesis of plasma-borne XIIIA is still not resolved (17). However, the follicle produces the monomeric form of factor XIIIA, and its regulated expression during follicle development suggests that it plays a role in this complex process.
To study this unexpected finding, we examined in close detail the expression pattern of follicular XIIIA during the entire sequence of follicle development, starting at phase 1 to the complete involution of the postovulatory follicle. Such an experiment is hampered by the fact that follicles taken from the last growth phase (F8 -F1) contain up to 15 g of yolk. RNA preparation from such follicles is very inefficient, and data obtained cannot be quantified appropriately. Thus, we have included follicles from phase 1 up to the smallest follicles at the beginning of phase 3 (F8 -F6), and representative postovulatory follicles covering the time period from after ovulation until almost complete resorption. Detailed evaluation on follicles during the last rapid growth phase will be presented in a separate experiment below. As shown in Fig. 5, transcripts for factor XIIIA are already present at the very beginning of follicle development, and the amount significantly increases until follicles reach the end of phase 2 (5-8 mm). This result confirms the initial finding from the differential mRNA display approach that led to the characterization of factor XIIIA in follicles. Apparently, mRNA levels decrease at the beginning of the rapid growth phase (approximately 8 mm) and further decrease after ovulation (Fig. 5, top). However, protein levels seem to increase during the last phase of follicle development, reach the highest levels shortly after ovulation, and stay high until the postovulatory tissue is completely resorbed (Fig. 5, bottom). This observation suggests that factor XIIIA is not catabolized during follicle development, but rather becomes deposited in the extracellular matrix of the follicle wall. To further evaluate this question and to pinpoint the site of synthesis of factor XIIIA in the follicle, we turned to the largest follicles (F4-F1) of the final growth phase. Despite problems to isolate RNA from follicles of this size, they can be dissected to separate and prepare the theca and granulosa cell layers free of most of the yolk (20). RNA and protein extracts were prepared from these tissues and subjected to Northern and Western blot analysis. The most striking result from this experiment was that expression of factor XIIIA in the follicle is restricted to the theca layer and virtually absent from the granulosa cell layer (Fig. 6A). Immunohistochemical analysis of the follicles showed that factor XIIIA is present in the theca externa and in the basement membrane but not in theca interna (Fig. 6B, Panels 1 and 2). Laminated staining of the theca externa, which is similar to that seen with anti-collagen-1 antibodies (Fig. 6B, Panel 3) and anti-fibronectin antibodies (Fig. 6B, Panel 4), together with the fact that granulosa cells do not express the protein, suggest that factor XIIIA is secreted from theca cells and deposited in the acellular matrix of the theca externa and possibly in the basement membrane.
To evaluate whether factor XIIIA is enzymatically active in the matrix of the follicle wall, an activity assay based on fluorescein-labeled cadaverine as substrate was performed on thin sections of the follicle. Background labeling of the sections was controlled by using EDTA (Fig. 7, A* and B*), which completely blocks the transglutaminase reaction. As seen in Fig. 7A, the   FIG. 5. Expression of factor XIIIA during follicle development. Total RNA (30 g/lane) was isolated from various follicles and postovulatory follicles and subjected to Northern blot analysis as described under "Experimental Procedures." Intensities of the signals were measured by densitometric scanning and normalized using the 18 S rRNA stained with Methylene Blue. Protein extracts (10 g/lane) obtained from the same tissues were subjected to Western blotting analysis using Ac16 as antibody as described under "Experimental Procedures." Relative amounts of factor XIIIA were normalized for total protein loaded. Relative amounts of transcripts and factor XIIIA protein are shown as open and filled bars, respectively. The schematic drawings of follicles of selected stages of maturation and postovulatory follicles are not to scale. distribution of the transglutaminase activity visualized by incorporation of the labeled substrate is identical with the distribution of the immunoreactive enzyme seen in Fig. 6B (Panels  1 and 2). Furthermore, the activity pattern is maintained in the early postovulatory follicle (Fig. 7B), when the highest amount of factor XIIIA is present in the follicle (Fig. 5). Note that at this stage the perivitelline membrane is lost (together with the ovulated oocyte) and the granulosa cell layer starts to disintegrate, whereas the theca externa reaches its largest extension. DISCUSSION Differential mRNA display screening for genes expressed during specific stages of chicken follicle development resulted in the identification of chicken factor XIIIA. The transglutaminase is expressed by cells of the theca externa and deposited in certain areas of the follicular wall. Usually, factor XIIIA is part of a large tetrameric complex consisting of two molecules of XIIIA and two molecules of XIIIB constituting plasma coagulation factor XIII. Upon activation by thrombin, factor XIII catalyzes the last step of the blood coagulation pathway by cross-linking soluble fibrin monomers to insoluble fibrin, which forms and stabilizes the final clot. Besides its abundant presence in plasma, factor XIIIA can also be found in certain cells like platelets (21), monocytes/macrophages (22), and some mononuclear cells in the skin (23). The cellular form of factor XIII, however, is a homodimer consisting of two A subunits FIG. 6. Expression of factor XIIIA in follicles during the rapid growth phase. A, thecae and granulosa cell layers from follicles during their rapid growth phase (F1-F4) were obtained, and total RNA and protein extracts were prepared as described under "Experimental Procedures." Total RNA (30 g/lane) derived from the tissues indicated were subjected to Northern blotting. Protein extracts (5 g/lane) from the same tissues were subjected to Western blot analysis using Ac16 as antibody as described under "Experimental Procedures." B, immunohistochemistry (panels 1-4) was performed on cryostat section of a chicken follicle (6 m) as described under "Experimental Procedures" using the indicated antibodies. Expressed factor XIIIA appears in green. Nuclei were counterstained with DAPI. Anatomically distinct parts of the follicles are indicated (TH ex, theca externa; TH in, theca interna; GC, granulosa cells; ૽, basement membrane; ଁ, perivitelline membrane). Panel 5 represents an image from a light microscope (differential interference contrast) of a section of a paraffin-embedded follicle (6 m).
only. Although the presence of factor XIIIA in these cells has been known for many years, the function of this form is still unclear. Recent reports (13,14,24) suggest that plasma factor XIIIA has specialized functions in certain other tissues and cells independent of its function in the blood. It seems to be associated with peripheral nervous tissue and may play a role in the regeneration of this tissue after injury (24). These authors present circumstantial evidence that XIIIA is at least partially synthesized by macrophages present in the nervous tissue. Another interesting aspect of a "non-plasma" form of XIIIA is directly related to the results presented here. Subtractive hybridization to isolate genes involved in hypertrophy of chondrocytes in the chicken lead to the identification of plasma factor XIIIA (13,14). During chondrocyte hypertrophy, factor XIIIA becomes up-regulated and externalized into the extracellular matrix, where it may serve to stabilize the matrix, which eventually undergoes calcification and the formation of bone structure. In the rat, however, expression of a tissue transglutaminase was described in terminally differentiating chondrocytes (25,26). Whether factor XIIIA expression in chondrocytes is specific for chicken and whether both transglutaminases serve similar or different roles remain to be established. In any case, factor XIIIA deposited in hypertrophic cartilage in the chicken is not derived from the plasma pool but is directly synthesized by the chondrocytes (14).
What is the function of factor XIIIA in the developing chicken follicle? The answer to this question may lie in the architecture of this continuously changing organ. The structural integrity of the follicle, which reaches a diameter of up to 3 cm and holds a yolk volume of approximately 15 ml at its maximal size before ovulation, is supported mainly by the mechanical strength of the theca externa. This layer is predominantly composed of sheets of fibroblast-like cells embedded in sheets of collagen fibrils (1) and fibronectin (27), giving this layer its stratified appearance (see Fig. 6B). The theca interna, however, contains many capillaries and cell clusters, which, due to their lipid granule content and rounded appearance, have been called thecal gland cells (28). These cells have been reported to be the main site of the production of progesterone and testosterone (29) and of alkaline phosphatase in the follicle (30). In sharp contrast to the theca externa, intercellular spaces of the interna are less well structured and contain only a few scattered collagen fibrils. Very little is known about the structure, composition, and function of the basement membrane that delineates the theca from the granulosa cell layer. However, immunohistochemical studies suggest that this structure contains the extracellular matrix protein fibronectin (27), apparently produced by granulosa cells (31).
The present immunohistochemical studies on the distribution of factor XIIIA in the chicken follicle (Fig. 6B) demonstrate that this protein is present in the theca externa and possibly in the basement membrane, although this could be demonstrated only with one of the antibodies used (compare Fig. 6A with 6B). Factor XIIIA is completely absent from the theca interna, the granulosa cell layer, and the perivitelline membrane. These results are supported by data obtained from expression studies showing that cells in the theca are the source of the protein.
The possibility that factor XIIIA is at least partially recruited from the plasma cannot be excluded at this point. However, it seems rather unlikely, because factor XIIIA in plasma always exists as a heterotetramer complexed to factor XIIIB. Because we could not detect any such complexes in the follicle (Fig. 3B), we suggest that factor XIIIA found in the follicle wall is produced by cells within the follicle, as is the case in the calcifying cartilage discussed above. As shown in Fig. 6B, the localization of XIIIA coincides with that of collagen-1 and fibronectin, both major extracellular matrix components of the theca externa. Because many extracellular matrix proteins, including collagen and fibronectin (32) are substrates for transglutaminases, we assume that factor XIIIA plays a central role in constructing and stabilizing the extracellular matrix of the theca externa. This may provide the mechanical strength necessary for progression through the rapid phase of oocyte growth.
This assumption also applies to the integrity of the postovulatory follicle, which after loss of its balloon-like tension by expelling the oocyte, does not disintegrate right away but stays metabolically active for several more days (2). There is another aspect about factor XIIIA function in the theca externa: It was recently shown that transglutaminases and especially factor XIIIA serve as substrates for cell adhesion (33). Upon inspec- tion of the ultrastructure of the theca externa (Fig. 6B), it becomes obvious that most of this tissue is a matrix, with only a few cells scattered in this layer. On the other hand, a massive presence of factor XIIIA in this tissue suggests that it serves other functions than catalysis of cross-linking of matrix proteins. For instance, factor XIIIA may be a true structural component of the matrix-like fibronectin, serving as an anchor for the adhesion of the fibroblast-like cells present in this stratum. Finally, follicular factor XIIIA may also serve the same function as in the blood. Ovulation of the follicle is started by rupture of the theca along the stigma, a special region of the follicle that is almost devoid of blood vessels. However, small capillaries are expected to rupture upon ovulation and factor XIIIA in the follicular wall might prevent microbleeding during this event.