| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print September 25, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Institut für Humangenetik und Anthropologie,* Universität Freiburg, Freiburg; the Institut für Humangenetik, Universität Göttingen, Göttingen; the Praxis für Humangenetik, Freiburg; the Abteilung Neuropathologie, Universitätsklinikum Göttingen, Göttingen; and the GSF National Research Center for Environment and Health,¶ Institute of Developmental Genetics, Neuherberg, Germany
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Sall1-null mutant mice die perinatally because of severe kidney dysgenesis or agenesis,14 whereas Sall2-deficient mice were reported to show no apparent phenotype.15 Homozygous Sall4 knockout mice die shortly after implantation, and Sall4 haploinsufficiency results in anorectal and heart anomalies as well as exencephaly.16 Mice lacking both Sall1 and Sall2 show kidney phenotypes comparable to those of Sall1 single knockout mice, suggesting that both genes do not genetically interact in vivo. In contrast, Sall1/4 double heterozygotes exhibit uni- or bilateral renal agenesis, exencephaly, anorectal malformations, and ventricular septum defects with increased incidence in comparison to Sall4 heterozygotes. This argues for a redundancy in the function of these two Sall genes.
The SALL proteins were shown to be important factors in complex regulatory networks in early and late vertebrate embryogenesis. Murine SALL4 is an essential transcription factor for the early development of inner cell mass-derived cell lineages and is required for the maintenance of embryonic stem cell pluripotency.17,18 At later stages, SALL4 controls Fgf10 expression in the limbs and Gja5 expression in the heart.19 In zebrafish, fgf10 expression is indirectly activated by sall1a and sall4 via activation of fgfr24.20 XsalF from Xenopus determines the position of forebrain and midbrain21 and Xsal3, csal3, and sall1a are essential for nephrogenesis.22-24
Cranial neurulation begins during gastrulation, when the neuroepithelium is induced to differentiate from the dorsal midline ectoderm to form the neural plate. Subsequently, the neural folds form at the edges of the neural plate, elevate from a semihorizontal to a vertical position, and shift from convex to concave in a complex process that involves midline neuroepithelial bending and expansion of the underlying cranial mesenchyme. Later on, the tips of the neural folds converge along the dorsal midline and fuse to form the neural tube.25-28 Neural tube defects (NTDs) are a group of heterogeneous and complex congenital anomalies of the central nervous system. Most of the mouse NTD mutants reflect a failure in neural fold elevation, and as a consequence, they present as exencephaly, spina bifida, or rachischisis at later stages. In mice, more than 150 mutations have been reported that cause NTDs, but no clinical or experimental study has provided unequivocal evidence for a definitive role for any of these genes in the causation of NTDs in humans. In addition, considerable variation is seen in the pattern of morphogenesis between inbred mouse strains. This indicates that neurogenesis is a complex process with a complex genetic network.
We have generated Sall2-null mutant mice, and in contrast to previous reports, we present evidence that Sall2 has a role in neurogenesis. The incidence of NTDs in Sall2–/– embryos is drastically increased in Sall2/4 compound mutants. In addition, compound Sall1/4 mutants also display NTDs similar to Sall2/4 mutants, suggesting an in vivo redundant role for the Sall genes during murine neurulation.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The targeting vector was generated by incorporating the 5' BglII/BglII fragment (4.3 kb) and the 3' BglII/BamHI (3.0 kb) fragment into the pTKneo vector (kindly provided by Nils Brose, MPI für Experimentelle Medizin, Göttingen, Germany). The 5' fragment was subcloned into pBluescript (Stratagene, La Jolla, CA) and subsequently cloned into a SalI-ClaI site 5' of the Neor gene. The 3' fragment was excised from a pBluescript subclone and inserted into a BglII-BamHI site 3' of the Neor gene. RI cells were electroporated, selected, and clones resistant to G418 were isolated and screened by Southern blot. The DNA isolated from the duplicate plate was digested with EcoRI, electrophoresed, transferred on nitrocellulose membrane (Amersham, Little Chalfont, UK), and hybridized to radioactive probes to confirm the correct homologous recombination. Nine of thirty-six clones were correctly targeted. Clones were then injected into 129SV/J blastocysts. Resulting chimeric animals were bred with C57BL/6 or 129SV/J females. Mutant animals studied were of F2 and later generations. Mice were genotyped by using genomic polymerase chain reaction (PCR). For genotyping we used the primer mix: SAF1: 5'-GGCTCACAACCATCCGTAACA-3'; ExR1 5'-ACACCTGCTCACCTCCATCG-3'; NeoR1 5'-GCGCGAATTCGATGATCCTGAACGGC-3'. The size of the amplicon is 400 bp for the wild-type allele and 620 bp for the Sall2 mutant allele.
Sall2+/– mice on a 129SV/J-DBA/2, 129SV/J-NZW, and 129SV/J-CD1 background were generated by crossing Sall2–/– mice on a 129SV/J background to mice on the inbred DBA/2 and NZW and the outbred CD1 background, respectively. Sall2–/– embryos from F1 intercrosses were analyzed at E12.5 to E14.5. Inbred NZW mice were obtained from The Jackson Laboratory (Bar Harbor, MA), and inbred DBA/2 and outbred CD1 mice were obtained from Charles River (Lyon, France).
Generation of Sall4+/gt Mice
Clone W097E01 (embryonic stem cells 129 x 1/SvJ2) carrying an insertion of the gene trap vector pT1βgeo in intron 1 of the Sall4 gene (Sall4GT(pT1Betageo)1Flo, hereafter termed Sall4gt) was identified within the German Gene Trap Consortium (GGTC; www.genetrap.de).29 Mutant mice were generated after injection into C57BL/6 blastocyst and resulting chimeras were bred to C57BL/6J females. Mice were genotyped using the primer mix: S4F 5'-TGGGGATTCCGGACTTGCTTC-3'; S4R 5'-TTTAAAAGCGGCGCCACTAGAG-3'; pGEOR6 5'-GAGATGGATTGGCAGATGTAGC-3'. The size of the amplicon is 284 bp for the wild-type allele and 234 bp for the Sall4gt mutant allele.
Generation of Sall1+/– Mice
The targeting vector was made by insertion of a 5' 6.2-kb XhoI/ClaI fragment and a 3' 1.1-kb XbaI fragment into the pTKneo vector. The construct was linearized with NotI and transfected into R1 embryonic stem cells. Clones resistant to G418 were isolated, screened, and analyzed as described above to confirm homologous recombination. Twenty-seven of forty-three clones were positive. Clones were then injected into 129SV/J blastocysts. Resulting chimeric animals were bred with C57BL/6 females. Mutant animals studied were of F2 and later generations. Mice were genotyped by using genomic PCR. For genotyping of the wild-type allele, we used the primers Ex2 F164 5'-TGATTACACGACATTTCCTACTG-3' and SA R164 5'-GACGACTCAAGTAAA AAGCACAA-3' (fragment size 400 bp). For the recombinant allele, PCR was performed with primers Neo F264 5'-GGCTGACCGCTTCCTCGTG-3' and SA R164 (fragment size 750 bp).
Histological Analysis
Embryos were collected in phosphate-buffered saline (PBS), fixed in Serra (ethanol:37% formaldehyde:acetic acid, 6:3:1), embedded in paraffin, and sectioned to 7 µm. Sections were stained with hematoxylin and eosin (H&E). For cell proliferation, sections were stained with rabbit anti-phospho-histone 3 (H3P) antibody (catalog no. 06-570, Upstate, Chicago, IL). Detection of apoptotic cells in paraffin sections was performed using a rabbit anti-caspase 3-active antibody (catalog no. AF-835; R&D Systems, Minneapolis, MN). The ABC kit (Vector Laboratories, Burlingame, CA) was used following the manufacturer’s protocol. In every section, the number of stained cells was scored and the surface area of neuroepithelium was determined using the AxioVs40 V 4.4.0.0 software (Zeiss, Jena, Germany). Images were captured by a Zeiss Axiocam MRc5 charge-coupled device camera.
In Situ Hybridization Analysis
Whole mount in situ hybridization was performed following a standard procedure,30
with digoxigenin-labeled antisense riboprobes for Sall2, Sall4,31
Fgf8,32
and Msx1.33
For expression analysis of Sall2 a 1082-bp Sall2 BamHI fragment of exon 2 (Figure 1A)
containing sequences coding for the third double zinc finger and 3'-UTR was subcloned into pBluescriptKS (Stratagene).
|
For co-localization assays, COS-7 and HEK293 cells were grown on coverslips (5 x 104 cells) and transiently transfected with expression plasmids using Lipofectamine (Invitrogen, Carlsbad, CA). After washing with PBS, the cells were treated with methanol for fixation (10 minutes), washed twice with PBS, and equilibrated in IF buffer [10 mmol/L Tris, pH 7.5, 100 mmol/L NaCl, 0,05% (v/v) Tween20, 1% (w/v) bovine serum albumin] for 30 minutes. Cells were incubated with primary antibodies (anti-FLAG; Sigma-Aldrich, St. Louis, MO; and anti-HA; BD Biosciences, Heidelberg, Germany), washed with PBS, incubated with secondary biotinylated antibodies (Vector Laboratories), and signals were detected using the fluorescent avidin kit (Vector Laboratories). Nuclear counterstaining was accomplished by 4,6-diamidino-2-phenylindole staining showing the location of the nuclei for orientation. Finally, coverslips were mounted on glass slides using Slowfade Gold (Invitrogen) and fluorescence signals were visualized by confocal laser-scanning microscopy.
Quantitative Real-Time Reverse Transcriptase (RT)-PCR
For quantitative real-time RT-PCR (qRT-PCR) total RNA was isolated from E8.5 stage embryos and cDNA was prepared by reverse transcription using oligo(dT) primers. The relative expression of the Sall4 exon 1-2 transcript was analyzed using the primers E1RTF1 (5'-GAAGCCCCAGCACATCAAC-3') and E2RTR2 (5'-CTGAGGCTTCATCGCAGTT-3'). The exon 1a-2 transcript was amplified with the primer pair E1aRTF1 (5'-CGGACTGCGACACACACC-3') and E2RTR2. The qRT-PCR was performed using the ABI Prism Real-Time Detection System 7900 (Applied Biosystems, Darmstadt, Germany). For the quantification of RT-PCR products, we used the SYBR Green assay system (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. In brief, the RT-PCR reactions were prepared by adding 5 µl of 2x SYBR Green mix, 2.5 µl of primer mix (forward and reverse, 1 pmol/µl), and 2.5 µl of cDNA template. The RT-PCR amplification condition was the following: 94°C 30 seconds, 63°C 30 seconds, 72°C 30 seconds for 45 cycles. Data analysis was done by SDS 2.1 software (Applied Biosystems) using standard curve method and Microsoft Office Excel (Microsoft, Redmond, WA). To verify the specificity and efficiency of real-time analysis, the PCR products were analyzed by monitoring their dissociation curves. The expression levels of investigated transcripts were normalized to β-actin mRNA expression to compensate for difference in sample amounts. The primer sequences used for amplification of the β-actin transcript were actinF (5'-CTTTGCAGCTCCTTCGTTGC-3') and actinR (5'-ACGATGGAGGGGAATACAGC-3').
Northern Blot
For expression analysis of Sall2 a 1082-bp Sall2 BamHI fragment34 containing sequences coding for the third double zinc finger and 3' UTR (exon 2) was subcloned into pBluescriptKS (Stratagene). A 32P-labeled probe of this subclone was generated by use of Rediprime II labeling system (Amersham Biosciences) and hybridized to Northern blots containing RNA of adult murine kidney tissue. RNA extraction, Northern blotting, and hybridization were performed as described.35
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
We have generated Sall2–/– mice by deletion of exons 1a and 2 via homologous recombination, resulting in a loss of all regions encoding zinc finger domains and loss of the complete coding region for the exon 1a-exon 2 transcript (Figure 1, A–C)
. Sall2+/– mice developed normally and had no phenotypic abnormalities. When intercrossed to generate homozygous Sall2–/– mutants, again no phenotypic abnormalities were observed on a mixed 129SV/J-C57BL/6 background (105 Sall2–/– embryos analyzed). However, when we analyzed Sall2–/– mice on a 129SV/J background, the litter size was constantly below the strain-specific average. We therefore decided to analyze the phenotypic appearance of the embryos and noted a significant perinatal lethality (data not shown). We observed abnormalities during neural tube development at E9.5 (Figure 1E)
and exencephaly at E11.5 in some Sall2–/– embryos (Figure 1G)
. To quantify the frequency of the phenotype, we collected 208 Sall2–/– embryos at E14.5, and 24 (11.5%) embryos displayed exencephaly (Figure 1I)
. While this work was in progress, a Sall2–/– mouse was described. Because these mutants developed to term without pathological findings, a dispensable role for Sall2 in murine embryogenesis has been postulated.15
In contrast, our results indicate that Sall2 is required for normal embryonic neural tube development, depending on the genetic background. To confirm this, we generated Sall2–/– mice on three new mixed backgrounds: 129SV/J-DBA/2, 129SV/J-CD1, and 129SV/J-NZW. We found NTDs in homozygous Sall2 mutants on the 129SV/J-CD1 and the 129SV/J-NZW backgrounds at a frequency of 12.9% and 17.1%, respectively, but not on the 129SV/J-DBA/2 background (Figure 1J)
. This result confirms that Sall2 plays a role during neurulation. It also shows that the frequency of NTDs displayed by Sall2–/– embryos varies depending on the genetic background, indicating that the different genetic constitutions of the various mouse backgrounds modulates the effect of the loss of Sall2.
Early Expression of Sall2 in the Cranial Region
Previous studies have revealed Sall2 expression in the metanephros and the spinal cord at E11.5. At later stages, the expression was localized to the kidney and the brain.15
Because we observed NTDs in Sall2–/– embryos already at E9.5, we decided to study Sall2 expression at early developmental stages. In situ hybridization for Sall2 revealed expression at E7.5 in the neural ectoderm, the mesoderm, and a broad anterior region of the head (Figure 2A)
. The same expression pattern was observed at E8.5 (Figure 2B)
. Later, the expression became restricted to distinct structures: strong signals were seen in the forebrain, midbrain, hindbrain, and the mesonephros at E9.5 (Figure 2, C–E)
and at E10.5 (Figure 2, F–H)
. Analysis of the sections indicated that Sall2 is expressed dorsally in the neuroepithelium all along the body axis (Figure 2, D and E)
. In addition, Sall2 expression could be detected in the otic vesicle at E9.5 (Figure 2, C–E)
, in the lens placode at E10.5, and in the optic cup at E11.5 (Figure 2, H and J)
. Starting at E11.5, Sall2 expression in the neural tube decreased but was maintained in the telencephalon and the third ventricle until E12.5 (Figure 2, I, K, and L)
. This pattern of expression is consistent with a role of Sall2 in neurulation.
|
The low penetrance of the cranial abnormalities in the Sall2–/– embryos might be explained by the compensation by other factors. Sall4, a closely related gene, was previously shown to display a similar expression pattern during embryogenesis,31
and we confirmed by in situ hybridization that both genes were simultaneously expressed in the developing neuroectoderm of the cranial region at E8.5 (Figure 3, A and B)
. To screen for a possible redundant role during neural tube development, we decided to generate compound Sall2/4 mutant mice.
|
. To determine the impact of the integration of the gene trap vector we performed quantitative RT-PCR at E7.5. Our analysis revealed that in Sall4gt homozygous mutants the mRNA expression initiated from exon 1 was completely abolished whereas the expression initiated from exon 1a was reduced to 45% when compared to wild-type embryos (Figure 3D)
. Therefore, the genetrap integration resulted in a hypomorphic Sall4 mutant allele. Analysis of the cross between Sall2+/–Sall4+/gt mice and Sall2–/– mice revealed that Sall2–/–Sall4+/gt developed NTDs with a full penetrance of the phenotype (24 of 24), whereas the phenotype was only partially penetrant in Sall2/4 compound heterozygous mutant embryos (4 of 23). We found no NTDs in the other genetic combinations (Figure 3I)
.
To determine when phenotypic differences between Sall2+/–Sall4+/gt and control embryos can first be detected, we examined H&E-stained sections of the cranial region between E8.0 to E8.5 (9 to 13 somite stage). At the 9-somite stage, the neural folds are already prominent in the cephalic region of both control and mutant embryos, lying in a semihorizontal position (Figure 3, J–M)
. At the 11-somite stage, the neural folds in the cephalic region are elevating and start to become parallel in control embryos (Figure 3N)
. This bending is also visible in the mutant neural folds, but to a much lesser extent, and the distance between the neural tips is always increased (Figure 3O)
. In a more rostral location, the development of an optic vesicle is visible in both control and mutant embryos (Figure 3, P and Q
; arrow), but the distance between the neural tips remains increased in mutant embryos (Figure 3Q)
. At the 13-somite stage, in control embryos, the neural folds in the cephalic region are lying parallel, and the neural tips are close to fusing (Figure 3R)
. In the mutant embryos, the degree of bending of the neural folds is similar to the 11-somite stage, and the neural tips are very far apart (Figure 3S)
. In a more rostral location, the edges of the neural tube are parallel and very close to each other in control embryos (Figure 3T)
, whereas they are very separated in the mutant embryos (Figure 3U)
. At E9.5, the neural tube is already fused in control embryos (Figure 3, V and X)
, whereas the neural folds remain unfused in the mutants (Figure 3, W and Y)
. Altogether, these results show that Sall2–/–Sall4+/gt embryos have a failure in neural fold elevation and fusion, which finally leads to the excencephally seen at later stages. They also indicate that Sall2 and Sall4 have a redundant function in neural tube closure.
Increased Apoptosis in Sall2–/–Sall4+/gt Embryos
To test whether a complete gene misregulation may be responsible for the phenotype observed in Sall2–/–Sall4+/gt embryos, we analyzed by in situ hybridization the expression of Fgf8, a marker for the neuroepithelium of the frontonasal region, the midbrain-hindbrain boundary, and the branchial arch,32
and Msx1, that is expressed in the forebrain and the midbrain.33
No difference in expression was seen for these two markers between Sall2–/–Sall4+/gt and wild-type embryos (Figure 4, A–D)
. Thus, a significant misregulation of these two marker genes seems not to be the origin of the neural tube closure defects seen in the Sall2–/–Sall4+/gt mutant embryos.
|
, top). In contrast, we found an increased number of apoptotic cells (Figure 4E
, bottom). Analysis of the neuroepithelium of control embryos stained for caspase 3 revealed some positive cells, in particular close to the bending of the neural tube (Figure 4, F, H, and J)
. In the equivalent regions in the mutants the number of positive cells was markedly increased (Figure 4, G, I, and K)
. Sall1 and Sall4 Act Redundantly during Murine Neurulation
The murine Sall1 gene is closely related to Sall2 and Sall4 and is also expressed in the neural tube. In compound Sall1/2-null mutants no neural tube abnormalities have been described.15 To screen for a possible redundant role between Sall1 and Sall4 in murine neurulation, we crossed Sall1+/– mice to mice harboring the hypomorphic Sall4gt allele. For this purpose we generated Sall1+/– mice by homologous recombination resulting in a deletion of exons 2 and 3 including all zinc finger domains. Heterozygous Sall1 mutant mice were phenotypically normal and fertile, whereas homozygous Sall1 mutants died perinatally as a result of a bilateral kidney agenesis, as previously described.14
By crossing Sall1+/– and Sall4+/gt mice we obtained 4 Sall1+/–Sall4+/gt mutants of 79 offspring. While this work was in progress, a Sall1/4 compound heterozygous mutant was described. Interestingly, 45% of these Sall1+/–Sall4+/– mutants showed NTDs, but none of them survived after birth.16 We therefore speculate that the remaining expression of our Sall4gt hypomorphic allele (see above) may account for the incomplete penetrance and therefore may rescue a small number of embryos from early lethality.
Analysis of the embryos from the cross between Sall1+/–Sall4+/gt mice and Sall1+/– mice revealed that Sall1–/–Sall4+/gt embryos presented with general growth retardation at E10.5 and died at approximately E11.5, probably because of cardiac failure (data not shown); they developed NTDs with a full penetrance of the phenotype (eight of eight) (Figure 5, E and F)
. In contrast the phenotype was partially penetrant in Sall1+/–Sall4+/gt compound heterozygous mutant embryos (9 of 20) (Figure 5, B, D, and F)
. We found no NTDs in the other genetic combinations (Figure 5F)
. In conclusion, our data suggest that the higher the dosage of Sall4 on a Sall1+/– background, or the higher the dosage of Sall1 on a Sall4+/gt background, the less severe is the penetrance of the resulting cranial phenotype. This supports the idea of functional redundancy of both genes during murine neurulation.
|
. |
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The low incidence of NTDs in Sall2–/– embryos, the similar structure of the SALL proteins, the overlapping expression patterns of Sall1, Sall2, and Sall4 in the cranial region, and the overlapping clinical features of patients with TBS and Okihiro syndrome led us to the hypothesis of a redundant function of the SALL proteins in neurogenesis. We therefore generated Sall2/4 and Sall1/4 compound mutants. Morphological analyses of the embryos revealed that the higher the SALL dosage, the less penetrant were the NTDs. In addition, we showed that SALL1, SALL2, and SALL4 co-localize in the nucleus, and previous GST-pulldown studies have demonstrated that the SALL proteins form homo- and heterodimers.40 These in vivo and in vitro results indicate an essential and redundant role of the murine Sall genes in neurulation.
Histological analysis of our Sall2/4 compound mutants revealed a failure in the elevation of the neural folds in the forebrain and midbrain. In Sall2–/–Sall4+/gt embryos the neural folds remain in a semihorizontal position and they never elevate and fuse. The exencephaly shown by mutant embryos at later stages is probably a direct consequence of the failure in elevation. Taken together, these results indicate that Sall2/4 have an essential role in the process of neural fold elevation during neurogenesis.
Neural tube closure is a complex and multistep process that involves lifting, bending, and fusion of the neural folds. Therefore, cell proliferation and cell death must be tightly regulated during this process. We did not find increased cell proliferation in Sall2–/–Sall4+/gt embryos, but we detected an increased number of apoptotic cells. Little is known about the molecular function of Sall2 and Sall4. Sall2 has been proposed to act in some cell types as a regulator of cell growth and survival,11 and Sall4 has been shown to be a regulator of survival and apoptosis in human leukemic cells.41 It is thus conceivable that both genes may have an anti-apoptotic function also during early neural development. Alternatively, Sall2/4 may control other unknown function(s) during neurulation, and the increased number of apoptotic cells may simply be a consequence of the failure in neural fold elevation.
NTDs are only rarely seen in TBS and Okihiro syndrome. In TBS, two families are known to us in which some affected members have Arnold-Chiari malformation type 1. In Okihiro syndrome, one patient with a SALL4 deletion was born with a meningomyelocele (J. Kohlhase, unpublished data). Therefore, NTDs are a rare feature associated with heterozygous SALL1 or SALL4 mutations in humans. However, it is conceivable that compound mutations in SALL1, SALL2, or SALL4 may result in defective neural tube closure also in humans. But such a severe phenotype would implicate embryonic lethality and hence escape the clinical diagnosis of TBS or Okihiro syndrome. Mutations in SALL2 have not yet been correlated with a human disease, and this may be because of the compensatory character of the other SALL genes. This assumption is strengthened by the fact that Sall2+/– mice are phenotypically normal and that Sall2–/– embryos exhibit a reduced penetrance of the NTDs (
10%) only in some mouse strains. It is therefore likely that the heterozygous loss of SALL2 can effectively be compensated by the related SALL proteins in human.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Deutsche Forschungsgemeinschaft (grants Ko1850/3-2 and Ko1850/6-2 to J. K. and Graduiertenkolleg Molekulare Genetik der Entwicklung to J. K.).
J.B. and A.B. contributed equally to this study.
Accepted for publication July 18, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
