Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Morphogenesis and differentiation of fetal organs is dependent upon precise signaling both within and between developing mesenchymal and epithelial cells. This signaling regulates cell proliferation, fate, migration, and differentiation (1). During development of the respiratory tract, embryonic cells organize along an axis (pattern formation) and differentiate such that the proximal structures (trachea) greatly differ from those of the more distal alveoli (2). As development proceeds, buds emerge laterally from the main bronchi, elongate, and branch. Temporal and spatial regulation of gene expression is essential to the polarization and outgrowth of distal buds in the rapidly branching lung epithelium. Although signal transduction in the developing lung is known to influence both branching and differentiation of the lung epithelium, the molecular mechanisms that regulate these events remain incompletely defined.

Early in development, nuclear import of transcription factors that translocate from the cytoplasm to the nucleus in response to hormones and growth agonists is critical to prominent signal transduction pathways that govern the branching process. Macromolecular traffic between the cytoplasm and the nucleus is an essential activity in eukaryotic cells (3, 4). Bidirectional nuclear-cytoplasmic transport through nuclear pore complexes (NPCs) is mediated by transport factors of the karyopherin- (kap-) family. Kap-s target import substrates by interacting with cognate nuclear-localization sequences (NLSs). The control of nuclear import of transcription factors represents a level of gene regulation integral to the developmental processes of both growth and differentiation (5).

In the mammalian lung, glucocorticoid hastens the maturation of type II epithelial cells. Glucocorticoid effects on lung development are likely to be mediated through regulation of the expression of multiple gene products.

To investigate molecular mechanisms of lung organogenesis, we searched for glucocorticoid-inducible genes in the developing lung in a fetal rat model. We applied the techniques of representational difference analysis (RDA) (8, 9), rapid amplification of complementary DNA (cDNA) ends (RACE) (10), and specific amplification of polymerase chain reaction (PCR) ends (system for PCR identification of cDNA ends; SPICE) (11, 12) to the identification of novel genes expressed in developing lung. Messenger RNA (mRNA) isolated from Day 20 fetal and adult rat lung was used to prepare two "representative amplicon" populations which were then subjected to successive rounds of subtractive hybridization/amplification to isolate target fetal lung- specific cDNAs. A single clone was used to isolate the 3.6-kb cDNA LGL2 (isolated from late gestation lung 2), encoding a deduced polypeptide (lgl2) of 963 amino acids. We investigated the tissue-specific expression, hormonal modulation, and cellular localization of LGL2 in fetal rat.

An initial comparison of lgl2 to sequences in the genome database ( BLAST search) suggested that lgl2 is a member of the kap- family of nuclear transport proteins. Kap- proteins have similar molecular weights (95 to 125 kd), isoelectric points (pI) (~ 5 to 5.4), and contain multiple HEAT repeats. By all these criteria, lgl2 was identified as a member of the kap- family. The developmental and hormonally modulated pattern of lgl2 expression in the pseudoglandular stage of development is coordinate with that of a number of key transcription factors that modulate formation of distal air sacs of the lung during this critical period.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Drugs and chemicals were obtained from the following sources: chromatographically purified collagenase (Type CLSPA) from Worthington, Freehold, NJ; culture media (minimal essential medium; MEM) and nylon (Hybond N) membranes from Amersham (Baie D'Urfé, PQ, Canada); penicillin, streptomycin, urea, agarose, TRIZOL, ethidium bromide (EtBr), Taq polymerase, and restriction endonucleases from GIBCO BRL Life Technologies (Burlington, ON, Canada); PCR primers from Sheldon Biotechnology (Montreal, PQ, Canada); Sequenase from Amersham; deoxynucleotides and RNA Guard RNase inhibitor from Pharmacia Biotech, Inc. (Baie D'Urfé, PQ, Canada); [³²P]deoxycytidine triphosphate, [³²P]adenosine triphosphate (ATP), and [³²P]uridine triphosphate (UTP) from Dupont Canada (Mississauga, ON, Canada); the Clontech PCR-Select cDNA subtraction kit and Marathon cDNA Amplification kit from Clontech (Palo Alto, CA); the ZAP-cDNA Synthesis kit from Stratagene (La Jolla, CA); Multiprime DNA and Enhanced Chemiluminescence Western Blotting kits from Amersham; and the Xpress System for Protein Expression from Invitrogen (Carlsbad, CA).

RNA Isolation

Total (nuclear and cytoplasmic) RNA was isolated by lysing the cells in TRIZOL. After extraction with chloroform, the RNA was ethanol-precipitated, collected by centrifugation, lyophilized, and dissolved in water. RNA integrity was confirmed by fractionation on 1% (wt/vol) agarose-formaldehyde gels, and staining the ribosomal RNA bands with EtBr. Poly(A)+ RNA was selected by passage through an oligo(dT) column.

RDA

RDA was carried out using the Clontech PCR-Select cDNA subtraction kit. A total of 2 µg of polyA+ mRNA prepared from total RNA isolated from Day 20 fetal (tester) and adult (driver) rat whole-lung tissue was reverse transcribed and used as template for second-strand synthesis. The tester and driver cDNA populations were separately digested with RsaI to obtain shorter blunt-ended molecules. The tester cDNA population was then subdivided into two portions and each was ligated with a different cDNA adaptor. Two hybridizations were performed. In the first hybridization, an excess of driver was added to each sample of tester. In this step the concentration of high- and low-abundance sequences is equalized, and there is enrichment of single-stranded molecules representing differentially expressed genes. In the second hybridization, the two primary hybridization samples were mixed without denaturation. Only the single-stranded tester molecules can reassociate and form hybrids with different adaptors on either end. Denatured driver was added (to remove remaining single-stranded sequences representing non-differentially expressed genes) before amplification with adaptor-specific primers. Molecules carrying adaptors at both the 5' and 3' ends represent differentially expressed tester sequences. Only these sequences were amplified. A second amplification using "nested primers" was carried out to eliminate background PCR products. After three successive rounds of subtractive hybridization/amplification, differentially expressed sequences (Figure 1a) were subcloned into pT-Adv for further analysis. A preparation of human skeletal muscle polyA+ mRNA provided by supplier (Clontech) was used as a control for all steps of the procedure (Figure 1a).

View larger version (60K): [in this window] [in a new window]   Figure 1.   (a) Agarose gel electrophoresis of difference products obtained from Day 20 rat fetal lung representative amplicons subtracted by adult lung. Lane 1, subtracted fetal lung cDNA; lane 2, unsubtracted fetal lung cDNA; lane 3, subtracted skeletal muscle cDNA; lane 4, unsubtracted skeletal muscle cDNA: lane 5, control subtracted skeletal muscle cDNA (provided by supplier). A 100-bp  ladder is included to indicate the size range of RNA fragments isolated in subtracted and unsubtracted fractions. (b) Reduction in abundance of G3PDH after subtraction/hybridization. PCR amplification reaction demonstrating absence of expression of G3DPH, a housekeeping gene, in subtracted fetal lung cDNA populations. Lanes 1-4, subtracted cDNA; lane 1, 18 PCR cycles; lane 2, 23 cycles; lane 3, 28 cycles; lane 4, 33 cycles. Lanes 5-8, unsubtracted cDNA; lane 5, 18 cycles; lane 6, 23 cycles; lane 7, 28 cycles; lane 8, 33 cycles. (c) Subcloning of RDA difference products. Fetal lung-specific cDNA probes subcloned from subtracted Day 20 fetal/adult lung cDNA populations after RDA. Arrow indicates clone used to isolate lgl2 (lanes 8, 9, and 10).

Assessment of Subtraction Efficiency

Measurement of the reduction of glyceraldehyde-3-phosphate dehydrogenase (G3PDH) cDNA (a housekeeping gene) in the subtracted fraction was used to assess the efficiency of subtraction (Figure 1b). The subtracted and unsubtracted fractions were amplified using G3PDH-specific primers. Aliquots were removed after 18, 23, 28, and 33 cycles of PCR amplification. Abundance of G3PDH was compared on a 2% agarose-EtBr gel.

Marathon RACE cDNA Amplification

A 610-base pair (bp) fetal lung LGL2 cDNA probe (Figure 1c) isolated by RDA was used to generate primers for 5' and 3' RACE reactions. Briefly, 1 µg of Day 20 fetal lung RNA was incubated with 20 U of avian myeloblastis virus reverse transcriptase (RT), 1 mM dNTP at 42°C for 1.5 h. Double-stranded cDNA (dscDNA) was synthesized by incubating the first-strand product in 0.5 mM deoxynucleotide triphosphate (dNTP) and second-strand cocktail at 16°C for 1.5 h. Thereafter, the dscDNA was blunt-ended using T4 DNA polymerase. The dscDNA was ligated to the Marathon cDNA adaptor (5'CCATCCTAATACGACTCACTATAGGGC3') with T4 DNA ligase, and used for 3'-RACE and 5'-RACE reactions. Nested primers for 3'-RACE were 5'CATGAGCCTGCCCACTTTCCCC3' and 5'GAGCAAATGGTTGAATGATG3'; nested primers for 5'-RACE were 5'CATAGGGAGGCAGTCATAC3' and 5'CTTGTCTTGTGGATCTGTTTC3'. PCR amplification conditions for adaptor and nested primers were: 94°C for 1 min; 5 cycles: 94°C 30 s, 72°C 4 min; 5 cycles: 94°C 30 s, 68°C 4 min; 25 cycles: 94°C 20 s, 65°C 4 min. The final PCR product (2.4 kb) was separated on a 1% agarose gel and the isolated band excised, subcloned, and sequenced.

SPICE

Northern analysis (discussed later with Figure 7a) revealed that the 2.4 kb cDNA sequence did not include the full-length cDNA of LGL2. The SPICE procedure (11, 12) was used to isolate the cDNA ends of LGL2.

View larger version (48K): [in this window] [in a new window]   Figure 7.   (a) Expression of LGL2 in fetal rat lung. Equal amounts of total RNA (30 µg) isolated from fetal whole lung at the indicated times were electrophoresed, blotted, and hybridized to a 2.4-kb LGL2 probe as described in MATERIALS AND METHODS. A probe for human 18S ribosomal RNA (rRNA) was used to control for loading, quality, and transfer of RNA. (b) Glucocorticoid induction of LGL2 expression in Day 20 fetal rat lung-adjacent fibroblasts. Equal amounts of total RNA (20 µg) isolated from Day 20 fetal rat lung cell cultures were electrophoresed, blotted, and hybridized to a 2.4-kb LGL2 cDNA probe as described in MATERIALS AND METHODS. Lung cell cultures were untreated or incubated in the presence of cortisol (107 M) for 24 h or dihydrotestosterone (108 M) for 48 h. A probe for human 18S rRNA was used to control for loading, quality, and transfer of RNA.

RNA from Day 20 fetal rat lung was used to create a cDNA library in the  ZAPII vector (Stratagene). A portion of this library was mass-excised with VCSM13 helper phage to yield a cDNA library in the pBluescript phagemid (Stratagene), using the protocol supplied with Stratagene's ZAP-cDNA synthesis kit. A total of 100 ng of this circular library was then used as a template for long-accurate, inverse PCR with Advantage KlenTaq Polymerase Mix (Clontech). The primers used for the first round of PCR were 5'TCTGTGTGTAGTGGGTGCACATGGCTCCTGAGC3' and 5'GGGGCACAGTTGCCTGTGAATGTTGGCCTCAAAC3'. The PCR product, consisting of phagemid plus additional cDNA sequences at either end, was visualized on an agarose gel as a band migrating at greater than 3 kb (the size of pBluescript). To increase the odds of obtaining a correct clone, a second round of PCR was performed with the nested primers 5'CACTCGTGGGGGTACGGGTAGGTGTAATCCTT-C3' and 5'TCTACCAGAGGACCTGAGCTCAGTTCCCACGAC3'. The PCR band was then excised from the gel, kinased, and self-ligated. Transformation of Escherichia coli DH5 with this DNA yielded a clone containing the complete phagemid sequence plus cDNA sequence both upstream and downstream from the PCR primers, joined together. This DNA was then autosequenced (Li-Cor model 4000L) from the T7 primer site in pBluescript.

Fetal Rat Lung Cell Primary Culture

Isolation and primary culture of the fetal rat lung cells was as described elsewhere (13, 14). Wistar rats of known gestational age (Day 0 = mating; term = Day 22) were obtained from Charles River (St. Constant, PQ, Canada), and killed with diethylether. The fetuses were immediately removed from the uterus and the fetal lungs dissected out. Epithelial and adjacent fibroblast (adjacent to the epithelium) cells were isolated from the fetal lungs as previously described (13, 14). Briefly, after trypsin dispersion, collagenase digestion, and several steps of differential centrifugation and adherence to plastic of adjacent fibroblasts, cells were incubated for attachment of epithelial cells. Nonadherent cells were removed from all cell cultures after overnight incubation. Cells were grown to confluence over 1 to 6 d in MEM + 10% fetal bovine serum, thoroughly rinsed in MEM (serum- and glucocorticoid-free), and then incubated in MEM alone. MEM controls were, in all other respects, handled exactly as the cells incubated in exogenous hormone (for glucocorticoid, cortisol 10⁷ M for 24 h; for androgen, dihydrotestosterone 10⁸ M for 48 h). An equal volume of the solvent in which the hormone was dissolved was added to the control medium. All experiments were performed 24 to 48 h after confluence. Viability and purity of the cultures were comparable to previously published data (14).

Hormonal Treatment of Fetal Lung Cells in Culture

Adjacent fibroblasts or distal airway epithelial cells in culture (Day 20) were exposed to glucocorticoid (cortisol, 10⁷ M, 24 h). We have previously shown that these doses modulate levels of glucocorticoid receptor mRNA and protein in the same model systems, fetal rat lung cells in culture, and lung tissue (15). mRNA levels were compared by Northern blot analysis in the presence or absence of factors added to the medium.

Sequence Analysis and Homology Modeling

lgl2 was subjected to a variety of sequence analysis tools. Database searches with the BLAST series of programs (16) revealed highest sequence similarity with members of the nuclear transport receptor superfamily. To identify homologies to additional members of the family, iterative profile searches were performed on the basis of the entire protein sequence. Sequences were aligned in ClustalW (17) and the relationship between proteins was viewed in Treeview (tree-drawing software by Roderic D. M. Page, 1998, at http://taxonomy.zoology.gla.ac.uk/rod/rod.html). Sequence data for the aligned proteins are available from GenBank/EMBL. We also analyzed lgl2 in PROPSEARCH, ProtParam, Psort II (identifies NLS sequences) and Psipred (secondary structure analysis), all accessed through the Proteomics tools at the ExPASy Molecular Biology Server at . SWISS-MODEL version 3.5 and the Swiss-PDB viewer version 3.5b (18) were used together for homology-based modeling of lgl2 (1-856) on the basis of the template structure 1QBKB.pdb (Chain B, Structure of the Karyopherin Beta2-Ran Gppnhp Nuclear Transport Complex). Both the first approach and the optimize mode were used. The generated model passed through all tests of ProModII for modeling and Gromos 96 for energy minimization.

Expression of Recombinant lgl2

The NH2 encoding terminus of the full-length cDNA of LGL2 was fused in frame to the "his" tag of the pRSET vector (Xpress system for protein expression; Invitrogen) and transformed in E. coli. Production of fusion protein was induced with isopropylthiogalactopyranoside. After lysis of the bacterial suspension, his-tagged protein was purified on ProBond resin (Invitrogen), and separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The enhanced chemiluminescence Western blotting kit (Amersham) was used to detect the recombinant lgl2 protein with antiXpress or T7 antibodies.

Northern Analysis

Total RNA (20 to 30 µg) was fractionated on a 1% agarose gel containing formaldehyde using 1 × 3-(N-morpholino)propanesulfonic acid buffer. The RNA was passively transferred to Hybond-N nylon membranes (Amersham) and ultraviolet crosslinked. A solution of 50% deionized formamide, 5× Denhardt's solution, 0.5% SDS, 6× saline sodium citrate (SSC), 0.1 mg/mL yeast transfer RNA, and salmon sperm DNA (0.2 mg/mL) was used for prehybridization at 42°C for a minimum of 6 h. Then hybridization occurred overnight at 42°C with 1 × 10⁹ cpm/mL of [³²P]UTP- labeled antisense cellular RNA (cRNA) probe in a solution that was of the same composition as for prehybridization except for the omission of SDS. Blots were then washed (highest stringency: 0.1× SSC, 0.5% SDS, at 65°C for 1 h) and transcripts were visualized after 48 h using standard autoradiography. To control for loading, quality, and transfer of RNA, the membranes were stripped and then reprobed with a [³²P]ATP-labeled RNA R18 oligoprobe to detect 18S RNA (GIBCO BRL). All Northern blots are representative of at least three independent experiments.

Fluorescence In Situ Hybridization

Metaphase chromosome spreads were hybridized to biotin-fluorescein isothiocyanate (FITC)-labeled BAC probes and developed with fluorescein-labeled avidin (Oncor, Tucson, AZ) according to the Oncor In Situ Hybridization Manual. Chromosome mapping was carried out at the MRC Genome Facility located at The Hospital for Sick Children (Toronto, ON, Canada).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of lgl2 in Fetal Rat Lung

We used RDA to identify genes expressed in fetal lung in late gestation. Adult lung cDNA amplicons were used as driver in successive rounds of subtractive hybridization/amplification to isolate target fetal lung-specific cDNAs. Figure 1a illustrates the difference in products obtained after three rounds of subtraction/hybridization. Subtraction of a preparation of human skeletal muscle mRNA, used as a control, is indicated for comparison. Efficiency of subtraction was assessed by comparison of the abundance of G3PDH (a housekeeping gene) before and after subtraction (Figure 1b). RDA products were subcloned into pT-Adv (Figure 1c).

A single RDA clone (610 bp, Figure 1c) that was conserved across species and had near-perfect homology to eight human/rodent expressed sequence tags was isolated for further investigation. The 5'- and 3'-RACE and SPICE reactions (see MATERIALS AND METHODS) were carried out to obtain the 3.6-kb cDNA LGL2, encoding a deduced polypeptide (lgl2) of 963 amino acids. The fetal rat LGL2 cDNA sequence is shown in Figure 2. LGL2 is the rat ortholog of KIAA599, a human cDNA sequence reported among a group of 100 unidentified human genes (19).

Expression of Recombinant lgl2 in E. coli

A his-tagged LGL2 cDNA (see MATERIALS AND METHODS) was used to transform E. coli. The immunoblot in Figure 3 illustrates the purification of recombinant lgl2 from E. coli cell extracts. A single band, at the expected size of 108 kd, was detected with the T7 antibody after affinity purification on Probond resin.

View larger version (59K): [in this window] [in a new window]   Figure 3.   Expression of recombinant LGL2 in E. coli. Recombinant LGL2 was used to transform E. coli. After lysis of the bacterial suspension, his-tagged protein was purified on ProBond resin (Invitrogen), and separated by SDS-PAGE. The enhanced chemiluminescence Western blotting kit (Amersham) was used to detect the recombinant lgl2 protein using the T7 antibodies. Lane 1, crude cell lysate; lane 2, precipitated proteins; lane 3, affinity-purified recombinant lgl2.

lgl2 Is a Novel Member of the Family of Nuclear Transport Proteins

A comparison of LGL2 with sequences in the genome database ( BLAST) suggested that lgl2 is a member of the kap- family of nuclear proteins, with closest sequence homology to TRN-SR (Figure 4). The amino acid composition as well as the predicted molecular weight (108 kd; kap- range, 95 to 125 kd), isoelectric point (pI = 5.3; kap- range pI ~ 5 to 5.4) and secondary structure of lgl2 clearly identified lgl2 as a kap- (Figures 4 and 5) (20). Members of the kap- family contain multiple HEAT repeats (structural domains predicted to be antiparallel, amphipathic -helices that form rod-like, tightly packed structures and present binding sites for additional proteins in loops between packed helices) (21). We generated a structural model of lgl2 by homology modeling based on the template structure of kap-2 (3) (Figure 5). The numbers indicate the HEAT repeats of kap-, which were confirmed in the crystal structure. The homology relationships of lgl2 with 29 members of the nuclear transport protein superfamily are illustrated in Figure 6.

View larger version (54K): [in this window] [in a new window]   Figure 4.   Alignment of lgl2 with TRN-SR. Asterisks indicate sequence identity; colons indicate conserved substitutions; and periods indicate semiconserved substitutions.

View larger version (36K): [in this window] [in a new window]   Figure 5.   Predicted structure of lgl2. A homology-based model of lgl2 on the template structure of kap-2 was generated in Swiss-Model. The numbers indicate kap-2 HEAT repeat regions (3).

View larger version (23K): [in this window] [in a new window]   Figure 6.   Homology relationships of lgl2 and nuclear transport receptors. lgl2 was initially analyzed in  BLAST (16) to identify sequence homologies with greatest significance (E-value threshold for inclusion in  BLAST iteration 1:0.001). To identify homologies to additional members of the family, iterative profile searches were performed on the basis of the entire protein sequence. Sequences were aligned in ClustalW (17), and the relationship between proteins was viewed in Treeview. Hs, Homo sapiens; Sc, Saccharomyces cerevisaie; Sp, S. pombe; Cs, Caenorhabditis elegans; Xl, Xenopus laevis; Dm, Drosophila melanogaster; Rn, Rattus norvegicus.

Developmental Expression of LGL2

To determine the developmental profile of LGL2 expression, we performed Northern analysis using a 2.4 kb rat LGL2 cRNA probe. The probe hybridized to a 3.6 kb mRNA species in fetal rat lung (Figure 7a). LGL2 was detected in the fetal rat in heart, kidney, brain, and intestine (Figure 8a). LGL2 was also detected at lower levels in all of these tissues in the adult rat (Figure 8b). Moreover, LGL2 mRNA was enriched in primary cell cultures of fetal lung distal airway epithelial cells when compared with adjacent fibroblasts (Figure 7b). LGL2 was detectible in rat whole lung at fetal Day 12 (not shown). The relative abundance of LGL2 mRNA in fetal lung was maximal during the pseudoglandular stage of development (Days 14 to 16) and decreased during the canalicular (Day 18) and saccular (Day 20, Figure 7a) stages. LGL2 could also be detected by Northern analysis in human lung, brain, heart, intestine, and kidney at fetal Week 16 (Figure 8c).

View larger version (39K): [in this window] [in a new window]   Figure 8.   Expression of LGL2 mRNA fetal (a) and adult (b) rat organs. Equal amounts of total RNA (20 µg) isolated from various fetal and adult rat tissues were electrophoresed, blotted, and hybridized to a 2.4-kb LGL2 probe as described in MATERIALS AND METHODS. A probe for human 18S rRNA was used to control for loading, quality, and transfer of RNA. (c) Expression of LGL2 in human fetal organs. Equal amounts of total RNA (20 µg) isolated from various human fetal tissues at Week 16 gestation were electrophoresed, blotted, and hybridized to a 2.4-kb LGL2 probe as described in MATERIALS AND METHODS.

Effect of Glucocorticoid on lgl2 Expression

To examine the effect of glucocorticoid on LGL2 mRNA expression, fetal lung fibroblasts and epithelial cells in culture were exposed to cortisol (10⁷ M) for 24 h. Cortisol increased expression of LGL2 in Day 20 fetal rat lung epithelial cells (Figure 7b). A lesser but detectable glucocorticoid stimulation of LGL2 expression was observed in Day 20 fetal lung-adjacent fibroblasts.

Chromosomal Mapping of LGL2

A 1-kb LGL2 cDNA probe, prepared by RT-PCR of human fetal lung (18 wk gestation) RNA was used to screen a BAC library. The isolated BAC clone was used as a probe for fluorescence in situ hybridization to map LGL2 to human chromosome 1p33-34.2 (Figure 9).

View larger version (59K): [in this window] [in a new window]   Figure 9.   Human LGL2 maps to 1p33-34.2. Metaphase chromosome spreads were hybridized to biotin-FITC labeled BAC probes and developed with fluorescein-labeled avidin. Arrows show location of genes on the chromosome.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We sought to identify glucocorticoid-inducible genes in developing lung. We cloned a novel gene, LGL2, that is enriched in lung epithelium and expressed in fetal brain, heart, intestine, and kidney. LGL2 is developmentally regulated, induced by glucocorticoid, and conserved across species.

Locally expressed molecules in lung mesenchymal and epithelial cells regulate branching and differentiation. These regulatory molecules are known to include transcription factors acting upstream (2, 22) as well as other growth and differentiation factors (22) that are involved in the complex network of reactions essential to terminal lung organogenesis. Until recently, the major approaches to identifying specific genes expressed during these processes were differential screening and subtractive hybridization. These approaches to the cloning and characterization of lung-specific genes have had limited success. We have therefore adapted the approaches of differential display (DD)-PCR (27) and RDA to the identification of novel genes expressed in developing lung in our fetal rat model. Using DD-PCR we identified LGL1, a glucocorticoid-induced, mesenchyme-specific gene maximally expressed in the rat fetal lung at gestational Day 21 (11).

More recently, we have applied the technique of RDA to our studies of genes important in lung development. Both DD-PCR and RDA offer advantages over previous screening approaches. Both procedures are highly sensitive, allow for the rapid identification of transcripts, and provide for the ability to process multiple RNA samples (30). RDA, however, is a method of subtractive hybridization coupled to amplification. Thus, whereas DD-PCR amplifies fragments from all represented mRNA species, RDA includes a subtraction step(s) that eliminates fragments common to two DNA populations. The major advantages of RDA over DD-PCR are: (1) the kinetic enrichment achieved by the subtractive process; (2) production of unambiguous difference products; (3) the use of full-length primers (24 mers) that avoids the mispriming which can occur with DD-PCR; and (4) the use of linkers ligated to full-length cDNA digests that allows for generation of amplicons representing the majority of cDNA species with a single primer pair. It is important to note that RDA can be "leaky." Multiple rounds of subtraction/hybridization before kinetic enrichment increases the likelihood of achieving a better representation of target-specific probes (8).

The use of global screening methods for gene identification requires a strategy for rapid identification of transcripts that are of greatest interest. LGL2 was chosen for further investigation on the basis of expected patterns of developmentally and hormonally modulated expression in fetal lung, and on significant protein homologies highly suggestive of analyzable protein functions.

The use of 5'- and 3'-RACE and SPICE procedures to isolate the full-length LGL2 cDNA allowed us to avoid the time-consuming effort of multiple library screens. An additional advantage to this approach is that these reactions allow for the isolation of sequences that are not limited by size. cDNA library inserts often do not exceed 2.0 kb. Although not all cDNA sequences identified by RDA will encode proteins that play either a regulatory or functional role in lung maturation, these cDNA sequences furnish a limited subset of cDNA probes.

Common motifs and significant protein homology constitute evidence suggesting functional similarities between proteins. Sequence homology searches have recently been used to identify novel kap (nuclear transport) proteins in a number of species. Comparison of rat lgl2 with sequences in the genome database (see MATERIALS AND METHODS) revealed significant homologies to nuclear transport receptor molecules in multiple species including yeast (Saccharomyces cerevisaie, S. pombe), Drosophila melanogaster, Caenorhabditis elegans, Xenopus laevis, and Homo sapiens (Figure 6). Members of the kap- family share limited amino acid sequence identity and considerable amino acid sequence similarity, and are most conserved at their amino termini which contains the Ran guanosine triphosphate binding domain. They have similar molecular weights and pIs, and contain multiple HEAT repeats (21). By all these criteria, lgl2 is identified as a member of the kap- family.

Kap- binding sites for diverse classes of transport substrates are found in the C-terminal half of their amino acid sequence (3). Kap-s target import substrates by interacting with cognate NLSs. Subsequent translocation through NPCs occurs in the presence of the guanosine triphosphatase, Ran. The transport of various NLS bearing substrates is mediated by different members of the kap- family. Proteins with classical NLSs, characterized by one or two stretches of basic residues, are generally imported by a kap-/kap- heterodimer. In this case, kap- binds the adaptor protein kap- which, in turn, binds the substrate NLS. Import pathways for other substrates involve a group of kap-s, all of which share limited sequence identity to kap-1, that bind substrates directly through specific NLSs. If lgl2 functions like kap-1 it is likely to interact with kap- but not directly with a substrate NLS. However, given that lgl2 shows greatest homology to TRN-SR and is more closely related to kap-s other than kap-1 (see dendogram, Figure 6), all of which do interact directly with specific NLS substrates, we believe it is likely that lgl2 will bind NLS-bearing substrates specifically and directly.

Early models of branching morphogenesis of the lung focused exclusively on the assembly and degradation of the extracellular matrix as major determinants of branching patterns (31). More recent evidence has emphasized the critical role of secreted signaling molecules in regulating mesenchymal-epithelial interactions that underlie the branching process.

Initiation of lung morphogenesis depends on coordinated transcriptional activation and repression involving a variety of signaling molecules including Sonic hedgehog, its receptor patched, and the transcription factors: hepatocyte nuclear factor (HNF-3), GLI, GLI2, GLI3 and Nkx2.1 (32). Mutant mouse embryos deficient in GLI2 and GLI3 exhibit lung hypoplasia, agenesis of the left lung, and abberrant lobulation of the lung (33). In the embryonic lung all three GLI transcription factors are strongly expressed during the pseudoglandular stage. Subsequent inductive events in the lung require epithelial-mesenchymal interaction mediated by specific fibroblast growth factor signaling and are modulated by growth factors (epidermal growth factor, platelet-derived growth factor, transforming growth factor-) and extracellular matrix components (31, 32).

The observed localization of LGL2 expression in time and space, and the glucocorticoid responsiveness of LGL2, are of great interest. LGL2 is hormonally modulated, enriched in epithelium, and maximally expressed in the pseudoglandular stage of development, concordant with the expression of key transcription factors that regulate the formation of distal air sacs of the developing lung. To our knowledge, this is the first report of a developmentally and hormonally modulated nuclear transport regulator in mammals.

The identification of lgl2 "cargo" substrates will likely provide important additional clues about its role in regulation of developmental signaling. Terminal branching of the fetal lung is an exquisitely modulated process that is likely to be dependent on multiple regulatory mechanisms. We speculate that lgl2 may be one important modulator of this process.