Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The distal respiratory epithelium of the vertebrate lung contains two anatomically distinct highly differentiated cell types. Type II alveolar epithelial cells are cuboidal cells that synthesize and secrete pulmonary surfactant (1). In contrast, type I alveolar epithelial cells are much larger cells with very thin cytoplasmic extensions; these cells cover more than 95% of the alveolar surface (2). During recovery from alveolar injury, type I cells slough from the epithelium and are replenished via a poorly understood differentiative process, with cells of type II origin (3, 4). In this scenario, a highly specialized, well-differentiated cell, synthesizing and secreting surfactant lipids and proteins, apparently shuts this system down, rearranges its cytoskeleton, and activates a new set of type I cell-associated genes. Recent cell culture studies lend strong support to the concept of a plasticity in the expression of type I and type II phenotypes by two cell populations (5, 6). For either biologic phenomenonlung injury or cell culture the molecular mechanisms underlying this transdifferentiation process are unknown.

A limited number of type I cell markers have been identified (7). Of these, the first described was rat type I cell 40-kD protein (RTI40), a 40-kD glycoprotein defined by a monoclonal antibody raised against partially purified type I cells (7). In the lung, RTI40 is localized to the apical plasma membranes of type I cells, which in turn cover more than 95% of the internal surface area of the lung. RTI40 is also expressed in the choroid plexus of the brain and ciliary epithelium of the eye; during fetal development, the RTI40 gene is widely expressed in embryonic brain and neural tissues in addition to the developing lung (11). A mouse RTI40 homologue, OTS8, was initially identified as an early response gene in a phorbol-ester-treated osteoblast cell line (12), also later independently characterized as gp38, a marker of the nonlymphoid stromal compartment of thymus, lymph node, and spleen (13). A canine homologue has also been reported (14). The RTI40 cDNA encodes a predicted 166 amino acid, largely extracellular protein with a single transmembrane domain and short cytoplasmic extension. Direct sequencing of purified rat lung RTI40 proteolytic peptides confirmed identity with the sequence predicted by the cDNA (15). The twofold difference between predicted (18 kD) and observed (40 kD) molecular weights likely reflects extensive O-linked glycosylation on the serine-, threonine-, and proline-rich (33%) extracellular domain, a property shared with epithelial mucins (16). Although at present there is no function known for RTI40, its striking anatomic location in a membrane covering more than 95% of the alveolar surface suggests an important function in lung homeostatic mechanisms.

We have undertaken a study of RTI40 in the context of its role as a type I cell differentiation marker. With a better understanding of RTI40 transcriptional regulation, we hope to identify mechanisms controlling the differentiated phenotypes of alveolar epithelial cells. We have utilized primary type II cells in a cell culture model that appears to replicate at least some aspects of this differentiation process. Freshly isolated type II cells do not express RTI40, but over the course of several days in culture come to express both RTI40 mRNA (Leland G. Dobbs, unpublished observation) and protein (7). Here we describe the structure of the rat RTI40 gene, define the start site of transcription, and assess the promoter activity of upstream flanking DNA in transient transfection assays.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation and Analysis of RTI40 Genomic Clones

A rat genomic DNA library (male Sprague-Dawley, Stratagene #945501) was screened with a partial RTI40 cDNA clone obtained from a rat lung expression library probed with an anti-RTI40 antibody (7) by Dr. James Fisher (Wayne State University, Detroit, MI). A subsequent 290 nucleotide (nt) 5'-rapid amplification of cDNA ends (RACE)- generated extension of this cDNA (17) was not included in our genomic library probe. Filters containing immobilized phage plaque DNA were hybridized overnight at 42°C in 0.8 M NaCl, 20 mM Pipes pH 6.5, 0.5% SDS, 50% formamide containing 100 mg/ml denatured fish sperm DNA and 3 × 10⁶ cpm/ml of ³²P randomly primed probe, washed at 50°C in 0.2× saline sodium citrate (SSC), 0.1% sodium dodecyl sulfate (SDS) and exposed to film. Positive clones were isolated to plaque purity and phage lysate DNA was prepared as described (18). Additional screens with upstream probes from positive clones were required to isolate the entire RTI40 gene. The phage clone 4RT1.2 was obtained by screening the library with a polynucleotide kinase labeled, sense strand oligonucleotide probe corresponding to the 5' end of the RTI40 cDNA (17), 5'-GGAGACATAAATTGCCG-3'); in this case, hybridization was at 40°C and washes were in 2× SSC, 0.1% SDS at 40°C. The set of positive phage clones was mapped by indirect end labeling with T7 and T3 probes of Southern blots prepared from partial restriction digests of phage DNA (Stratagene protocol manual). Manual sequencing (19) (Sequenase; USB, Cleveland, OH) was performed on overlapping clones generated by exonuclease III digestion (20) using the Erase-a-base kit (Promega, Madison, WI). Sequences were entered, aligned, and analyzed with GeneWorks (Intelligenetics, Mountain View, CA) and the Wisconsin Sequence Analysis Package (GCG, Madison, WI) software programs.

S₁ Nuclease

The Berk-Sharp method was used as modified by Ambion (Woodlands, TX) (21). Total RNA (22) from adult rat lung and from type II cells cultured for 5 d was hybridized with a single-stranded DNA probe generated from the plasmid p4RT1.5 (for description, see PLASMIDS). The NotI linearized plasmid was denatured and hybridized with a primer antisense to nucleotides 184 to 201 of the RTI40 transcript (5'-TCTTGATCTCGTTGGAGC-3'; the initiator methionine codon starts at nt 201). The primer was extended with Klenow fragment in the presence of ³²P-dCTP and the 336 nt product (including 36 nt of polylinker sequence) was gel-purified. A sequencing ladder, generated with the same primer-plasmid combination described previously, was used to determine the precise endpoint of the protected fragment.

Cell Culture and Transfection

NCI H441A cells (23) were cultured in Dulbecco's modified Eagle's medium (DMEM) H-21 supplemented with 5% fetal bovine serum. Transfection was by a modification of the adenovirus-assisted diethylaminoethyl (DEAE)-dextran method of Forsayeth and Garcia (44). Two microliters of DEAE dextran (8 mg/ml water) were added to 50 µl adenovirus-infected 293 cell lysate (for preparation, see Forsayeth and Garcia [44]); this was then added to 15 µl of 1 mM Tris pH 7.5, 0.1 mM EDTA containing 5 µg reporter plasmid and 0.5 µg p6RL (a respiratory syncytial virus promoter-driven -galactosidase vector). The DNA-adenovirus-DEAE-dextran mixture (67 µl) was added directly, without change of media, to H441 cultures plated 1 d previously with 2 × 10⁵ cells in 2 ml media. After the cultures were incubated for 1 h at 37°C, the supernatant liquid was aspirated and the cells were washed twice with 1 ml of phosphate-buffered saline (PBS) prior to incubating an additional 48 h in complete media. Extracts from harvested cell pellets were prepared in 150 µl Reporter Lysis Buffer (Promega). Luciferase and -galactosidase enzyme activities were determined with luminescent substrates as recommended by Promega and Tropix (Bedford, MA), respectively. The value for luciferase activity in each sample was divided by the corresponding -galactosidase activity to correct for variations in transfection efficiency.

Rat type II cells were isolated and cultured as described previously (24). For transfection, 1 × 10⁶ cells/35 mm tissue culture dish were cultured for 2 d; the cells were washed twice with PBS, then incubated at 37°C for 1.5 h with 1.0 ml transfection mix prepared by adding, in order, to 956 µl DMEM H16: 50 µl adenovirus-infected 293 cell lysate (see above), 10 µl DEAE dextran (8 mg/ml water), and 4 µg reporter plasmid in 4 µl 1.0 mM Tris, pH 8.0; 0.1 mM EDTA. Following transfection, the cells were washed twice with PBS and incubated an additional 2 d in complete medium prior to harvest, extract preparation, and luciferase measurement as described previously for H441 cells. In these experiments, luciferase activity was expressed per microgram of extract protein because we were unable to detect p6RL- dependent -galactosidase activity above the high background of endogenous enzyme activity in type II cell extracts.

Plasmids

Plasmids were grown in Escherichia coli DH5 and purified on Qiagen columns (Chatsworth, CA). The RTI40 -phage inserts were subcloned as EcoRI fragments into pBluescript SK (Stratagene), then further subcloned as necessary using available restriction sites. Of particular relevance are two adjacent 6.5 and 2.4 kb EcoRI fragments from 4RT1.4 (Figure 1a) containing the RTI40 transcription start site and 5' flanking region, respectively; these subclones are designated p4RT1.4 and p4RT1.2. Subclone p4RT1.5 contains the transcription start site on a 1.8 kb EcoRI to XbaI fragment subcloned (including polylinker sequence) as an XbaI fragment from p4RT1.4. Subclone p4RT1.9, containing the RTI40 transcription start site and upstream sequence to 2.5 kb, includes the 204 bp SstI to EcoRI fragment of p4RT1.5 ligated with the 4.4 kb NcoI to SstI and 1.4 kb NcoI to EcoRI fragments of p4RT1.2. Plasmid pOPLO.R1, a luciferase reporter gene containing nucleotides 99 to +103 of the RTI40 promoter, was constructed by inserting a PstI-SstI fragment from p4RT1.5 and an 860 bp SstI-SphI fragment from pOALO 67R (a human SP-A2 promoter-luciferase plasmid (25); fragment contains only luciferase and polylinker) into the 4.3 kb SphI-PstI fragment of pFOXluc (26) (a gift of Dr. Michael German, University of California at San Francisco, San Francisco, CA). To construct pOPLO, a 2.4-kb BglII-SalI fragment from p4RT1.9 was inserted into BglII and SalI digested pOPLO.R1. The 5'-deletion mutants (Figures 4-6) were generated by Henikoff deletion (20) of KpnI-SalI-digested pOPLO. The distal region internal deletion resulted from religation of AvrII-digested pOPLO. The Drosophila melanogaster adh promoter-luciferase reporter plasmids are based on pODLO.F (Figure 5, construct 1), constructed by ligating HindIII-BstEII fragments of pODLO (750 bp) (27) and pFOXluc (4.5 kb). The distal element plasmids pPDLO.AvrIIF and pPDLO.-AvrIIR (Figure 5, constructs 2 and 5) contain opposite orientations of the 452 bp AvrII fragment from pOPLO in the SpeI site of pODLO.F. The proximal element was isolated as a BclI-BglII fragment from the pOPLO-163 deletion endpoint (Figure 4a) and inserted in both orientations at the polylinker BglII site in pODLO.F to create pPDLO.-BBF and pPDLO.BBR (Figure 5, constructs 3 and 6). The four proximal-distal element combination plasmids (Figure 5, constructs 4, 7, 8, and 9) were constructed by inserting the proximal element (see previous discussion) into BglII-digested pPDLO.AvrIIF and pPDLO.AvrIIR; these plasmids are designated pPDLO.ABFF (construct 4), pPDLO.ABRR (construct 7), pPDLO.ABRF (construct 8), and pPDLO.ABFR (construct 9) (see Figure 5 for orientation). Plasmid constructs were verified by a combination of diagnostic restriction digests and DNA sequencing.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Structure of the rat RTI40 gene. (a) The diagrammed nine overlapping cloned lambda phage inserts isolated with RTI40 cDNA and oligonucleotide probes (top). Vertical marks indicate EcoRI sites. Intron (solid line) and exon (open boxes) organization of the RTI40 gene are shown (center). The sequenced regions are marked with a wide solid line. Asterisks indicate positions of ID repetitive elements; the unsequenced regions have not been examined for additional ID elements. (Bottom) Restriction map of the region. B = BamHI; E = EcoRI; H = HindIII. (b) Exon/intron borders of the RTI40 gene as determined from the sequenced region. Exons are in uppercase and introns are in lowercase. The nucleotide number of the cDNA sequence (17) at each junction is indicated above the sequence, and the corresponding amino acid residue and number is indicated below. The length of intron 1 is approximate based on restriction fragment mapping.

View larger version (41K): [in this window] [in a new window]   Figure 4.   Deletion analysis of the RTI40 promoter in H441 cells. (Top panel ) 5' Nested deletion series. H441 cells were cotransfected with the indicated luciferase reporter gene plasmid containing RTI40 upstream sequence terminating at the designated endpoint and p6RL, an RSV--galactosidase internal control plasmid. Luciferase activity, assessed on Day 2 after transfection, is expressed relative to that of the promoterless control vector pFOXluc (26). Each error bar represents the standard deviation of the means of four independent experiments, each assayed in triplicate. (Bottom panel ) Internal deletion of the distal element. The AvrII fragment between nucleotides 1247 and 795 was deleted from the full length RTI40 promoter. The resulting plasmid and the indicated 5' deletion constructs were compared in H441 transfection assays as in the top panel.

View larger version (40K): [in this window] [in a new window]   Figure 5.   Function of the proximal and distal RTI40 elements on a heterologous minimal promoter. Proximal (solid box) and distal (cross-hatched box) elements were fused to the drosophila adh promoter (shaded box) at position 33 from the transcription start site (small arrow). The orientation of these elements with respect to the intact RTI40 promoter is indicated by the heavy arrows. Plasmids were transfected into H441 cells and assayed as in Figure 4. Luciferase activity for each construct is expressed relative to that of the minimal adh promoter alone (construct 1).

View larger version (26K): [in this window] [in a new window]   Figure 6.   Deletion analysis of the RTI40 promoter in cultured type II cells. Rat type II cells were transfected 2 d after plating with the indicated RTI40 5' deletion constructs. Luciferase activity was determined 2 d after transfection and is expressed relative to the promoterless control vector pFOXluc. Each error bar represents the standard deviation of the means of three experiments with different preparations of type II cells with triplicate samples in each experiment. *Indicates statistical significance versus the next upstream deletion with P  0.025 as determined by Student's t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Structure of the RTI40 Gene

We screened a rat genomic DNA library with a partial RTI40 cDNA probe (kindly provided by Dr. James Fisher) and obtained several clones encoding downstream RTI40 exons. This and subsequent screens with progressively more 5' probes, including one screen with an oligonucleotide corresponding to the extreme 5' end of the published RTI40 cDNA sequence (17), generated the set of nine overlapping phage clones diagrammed in Figure 1a. The partial DNA sequence of this region (GenBank accession number U92440) revealed a 35-kb RTI40 gene containing six exons and five introns (Figure 1b). Exon 1, encoding the 5'UTR and a 22-amino acid peptide leader segment, is separated from the remaining five exons by a 25-kb first intron. The next 105 amino acids, constituting the predicted extracellular domain of RTI40 (amino acid 23 through 128), are encoded by exons 2, 3, and 4. The peptide predicted by exon 3 is particularly rich (43%) in serine, threonine, and proline making it a likely target for O-linked glycosylation (28). In addition, there are several potential O-linked glycosylation sites in exons 2 and 4. This region is followed by exon 5, which encodes 38 amino acids of a putative transmembrane spanning and cytoplasmic domains. Finally, exon 6 encodes the C terminal proline, the TAA stop codon, and 1,114 nucleotides of 3'-UTR. Examination of the RTI40 gene sequence revealed six separate occurrences of the rat Identifier element (ID), a family of short interspersed elements, or SINEs, repeated 130,000 times in the rat genome (29). One ID element occurs in the 5'-flanking region of the RTI40 gene, whereas the remaining five are located within introns (Figure 1a).

RTI40 Transcription Start Site and 5' Flanking Region

To map precisely the initiation site of RTI40 transcription, we performed an S₁ nuclease protection assay (21) on RNA from adult rat lung and day 5-cultured type II cells (Figure 2). A labeled antisense single-stranded DNA probe starting at the initiator methionine codon in the cDNA sequence was extended to an EcoRI site in the putative RTI40 promoter. Following hybridization of the probe to RNA, S₁ nuclease-resistant products were analyzed on a denaturing polyacrylamide gel. RNA from either rat lung or cultured type II cells yielded a somewhat diffuse band migrating at approximately 200 nucleotides. Further analysis on a longer gel revealed a cluster of fragments extending from 201 to 204 nucleotides. This heterogeneity most likely reflects incomplete S₁ nuclease digestion at the single-stranded/ heteroduplex junction (30), although the possibility of multiple transcription start points cannot be eliminated. Based on these results, we designated the transcription start point as the adenosine residue at the 5' end of the shortest S₁ nuclease-protected fragment (201 nt). The sequence of this region (TTGCCA₊₁GTTG) matches a consensus eukaryotic polymerase II initiator element (PyPyA₊₁NT/APyPy) at all but the last nucleotide (31, 32).

View larger version (54K): [in this window] [in a new window]   Figure 2.   Analysis of the RTI40 transcription start point by S1 nuclease assay. (a) RNA from adult rat lung (lane 1), Day 5 cultured type II cells (lane 2) or yeast RNA (lanes 3 and 4) was hybridized with a 32P-labeled probe overlapping the putative 5' end of the RTI40 mRNA (diagrammed in panel c). Samples in lanes 1-3 were digested with S1 nuclease; lane 4 was left undigested to visualize the free probe. All samples were analyzed on a denaturing polyacrylamide gel along with a sequencing ladder (lanes G, A, T, C), primed with the identical primer used to generate the probe. (b) A small residual sample from adult rat lung (lane 1 in the left panel) was electrophoresed for a longer time on a second polyacrylamide gel along with the same sequencing ladder (G, A, T, C), permitting better resolution of the transcription start point (indicated with a dot on the right-hand side of the sequence). Note that the sequence shown is of the noncoding (bottom) strand. (c) Schematic of the experiment. The 32P-labeled single-stranded DNA probe extends from the initiator methionine codon (ATG) to the EcoRI site (RI). RNA (squiggle line) was hybridized with the probe and digested with S1 nuclease to yield the protected fragment. The arrow denotes the position of the 5' end of a previously published RACE-generated cDNA (17).

The sequence of the putative RTI40 promoter region, extending from 2496 through the end of exon 1 is presented in Figure 3. Inspection of the upstream sequence revealed a TATA homology, CTTAAATTGCCG, starting at 31 (31) and a GC-rich region between 41 and 62 that includes several putative binding sites for the Sp1 family of transcription factors (33). The combination of S₁ nuclease protection data, the TATA-like homology, and GC-rich sequence strongly suggested an RTI40 promoter in this region of DNA. In addition, a DNase I hypersensitive site (34) appears in the chromatin of this region as RTI40 expression commences in cultured type II cells (Jeff N. Vanderbilt, manuscript in preparation).

View larger version (76K): [in this window] [in a new window]   Figure 3.   Nucleotide sequence of RTI40 exon 1 and 5'-flanking region. The arrow indicates the transcription start point (+1) and asterisks underscore the TATA homology; putative binding sites adjacent to the TATA box are indicated for TTF1 (broad overline) and Sp1 (thin underlines). The dashed underline denotes the sequence of a highly repetitive rat identifier element (29). The amino acid translation of exon 1 is shown above the sequence.

Studies of RTI40 Promoter Activity

We tested the genomic DNA surrounding the transcription start site identified in Figure 2 for functional promoter activity by transiently transfecting either H441 cells or cultured type II cells with promoter-reporter gene constructs. H441 is a lung adenocarcinoma cell line frequently used for expression studies of lung-specific genes (23, 35- 37). Type II cells, either in situ or freshly isolated, do not express either mRNA or protein for RTI40. As type II cells are cultured on plastic substrates, they express progressively increasing amounts of both RTI40 mRNA (Leland G. Dobbs, unpublished observations) and protein (7).

We first ligated the 2496 to +104 RTI40 promoter fragment to a luciferase reporter gene. When this plasmid, designated pOPLO, was transfected into H441 cells (2496 endpoint in Figure 4a), there was a 34-fold increase in luciferase activity over the control vector pFOXluc, which contained no promoter sequence. Using 2496 of the RTI40 promoter as an entry point, pOPLO was digested with exonuclease III to generate a series of 25 nested 5' deletions. These plasmids were tested in H441 cells to identify RTI40 upstream regions that contribute to promoter function. For the 12 plasmids with 5' deletion endpoints between 2496 and 1143, there were slight variations in luciferase activity. However, the next deletion, to 976, produced a significant fall in activity. Luciferase activity then remained constant for eight different deletion endpoints between 976 and 254 bp. The sharpest and most extensive drop in activity occurred for deletions between 254 and 41 bp with activity falling from 31-fold to only 1.6-fold above control over this span. Intermediate endpoints in this region at 163, 100, and 81 produced 19, 6, and 3.5 times control levels. There was no significant luciferase activity from reporter plasmids containing reverse orientations of the 2224, 1771, 587, or 100 fragments (data not shown). These transfection experiments define two regions important for RTI40 promoter activity, one between 1143 and 976 bp and the other between 163 and 81.

To test the functional significance of the RTI40 TATA box homology starting at position 31, we generated a four-base-pair mutation in the wild type sequence (CTTAAA changed to GTCGAC). The mutation was introduced into the 1143 bp, 254 bp, and 100 bp RTI40 promoter reporter constructs then tested for effects on promoter activity in transfected H441 cells. For each different endpoint, the TATA box-site-directed mutation caused a 75% decrease in luciferase expression compared to the respective wild-type control plasmid (data not shown). This sequence appears critical for RTI40 promoter activity, perhaps reflecting activity of the TATA binding protein TBP. A closer examination of the surrounding sequence revealed a potential thyroid transcription factor 1 (TTF1) binding site (8 out of 10 bp match) positioned immediately 5' of the TATA box and 3' of the Sp1 sites mentioned previously (please see Reference 38 for a review of TTF1 and Sp1 effects on lung specific promoters). Interestingly, the Clara cell secretory protein (CCSP) promoter contains a similar positioning of TTF1, Sp1, and TATA elements (39). Future experiments will examine these putative elements in the RTI40 promoter for functional relevance.

The more distal of the two RTI40 promoter elements was less potent than the proximal element in the H441 transfection experiments (1.7-fold versus 19-fold effects, respectively). To characterize the distal element further, we created an internal deletion of nucleotides 1247 to 795 from pOPLO (Figure 4b). Promoter activity of this construct was reduced by 80% in H441 cells compared with pOPLO. In contrast, the 5' deletions to 1143, 976, and 876 caused, at most, a 50% reduction in luciferase activity, suggesting that the distal element may interact with regions further upstream. We conclude that the distal element is important for full activity of the 2.5 kb RTI40 promoter fragment.

The two promoter regions were then tested singly and in combination for their ability to activate a linked minimal eukaryotic promoter containing only a TATA box and initiator element. Distal (1247 and 795) and proximal (163 to 81) elements were fused at nucleotide 33 of a Drosophila alcohol dehydrogenase promoter-luciferase reporter gene combination (40), and the resulting plasmids were transfected into H441 cells. The distal element alone doubled adh promoter activity, whereas the proximal element alone activated 7.8-fold (Figure 5, constructs 2 and 3). In combination, these elements activated 14-fold (Figure 5, construct 4). The elements also activated transcription when ligated to the adh promoter in the reverse orientation (Figure 5, constructs 5-7) and when combined in opposite orientations (Figure 5, constructs 3, 5, 8, and 2, 6, 9). In all four combinations tested, the effects were greater than the additive sum and approached the multiplicative product of the two elements' separate activities. These results strongly suggest a cooperative interaction between the two elements. For reference, the region between 77 and 1771 activated the adh promoter in these experiments by 23-fold and the native RTI40 promoter, containing sequences from 1771 to +104, activated luciferase expression 32-fold over control levels (data not shown).

Replacement of type I cells in a damaged alveolar epithelium is thought to occur from a population of type II cells that flattens, increases in surface area, ceases production of surfactant, and assumes the morphologic characteristics of type I cells (3, 4). Freshly isolated type II cells, to some extent, undergo a similar process during the first few days of culture on plastic substrates. Over the same time course, these cells progressively increase expression of RTI40 mRNA and protein (7), thereby suggesting that this model may be appropriate to study RTI40 gene regulation. We transfected a subset of the deletion mutants tested in Figure 4a into type II cells two days after plating and harvested the cells 2 d later. As shown in Figure 6, RTI40 promoter activity was constant for the deletion endpoints tested between nucleotide 2497 and 163; further deletion to 100 resulted in a marked drop in promoter activity. The luciferase activity measured in these transfected cultures was somewhat weaker and more variable than that of H441 cells, most likely because of the low transfection efficiency of type II cells. To date, this variability has precluded a detailed characterization of the distal element in this system; however, the effects of the proximal element, at least for those endpoints tested, appear similar for H441 cells and cultured type II cells.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have isolated overlapping phage genomic clones that define the gene for rat RTI40, an integral apical plasma membrane protein localized within the lung to type I alveolar epithelial cells, a cell population that covers more than 95% of the internal surface of the lung. The gene for this comparatively small protein (166 amino acids) contains six exons and five introns dispersed over 35 kb of rat DNA. The time required for RNA polymerase to transcribe genes of this large size could have several biologic consequences (41), notably, a significant lag between changes in transcription initiation rate and corresponding changes in mRNA level. The first intron occupies well over 70% of the gene; its 25-kb length places it among the longest 5% of characterized introns (42). Such a long intron also raises the possibility of undetected exons that could be alternatively spliced into the RTI40 transcript. In addition, the organization of exon 6, encoding a stop codon after only one amino acid, suggests exon skipping as a means to produce an alternative protein containing a longer cytoplasmic domain. At present, we have no direct evidence for RTI40 alternative splicing; however, a more extensive analysis of the RTI40 mRNA sequence in lung and other tissues expressing the protein will be required to address this possibility.

S₁ nuclease protection assays mapped the RTI40 transcription initiation site to a position 31 base pairs downstream of a TATA homology and provided no evidence for additional initiation sites upstream to the limit of the probe at 100 bp. Analysis of a 2.5-kb fragment of RTI40 5'-flanking DNA by deletion mutagenesis revealed two regions that contribute to promoter activity in transfected H441 cells, a distal element between 1247 and 795 and a proximal element between 163 and 81. In addition, we showed that site directed mutation of the TATA box homology reduced transfected reporter gene expression by 75%. Both proximal and distal elements activated transcription when positioned upstream of a heterologous, minimal eukaryotic promoter. Our results suggest that the two elements interact cooperatively because the proximal and distal elements together activate transcription to a greater extent than the sum of their activities when tested individually. Furthermore, both forward and reverse orientations of either element were effective in these experiments even when positioned out of the normal context for each element. Position and orientation independence in the activation of a linked promoter are properties of a classical enhancer element (43). Thus, our results are consistent with the hypothesis that both proximal and distal elements are enhancers, although additional studies will be required to support this conclusion firmly.

Transfected RTI40 promoter constructs were also active in type II cells maintained in vitro on tissue culture plastic. This system would appear well suited for studies of RTI40 gene regulation because freshly isolated type II cells, initially negative for endogenous RTI40 mRNA and protein, become positive within the first 24 h of culture (Reference 7 and unpublished observations). Type II cells are difficult to transfect by standard techniques. We utilized an adenovirus-assisted DEAE-dextran approach (44) to deliver conventional plasmid based reporter genes to both H441 and type II cells. The transient transfection studies in type II cells were more variable than their H441 counterparts; this may have prevented detection of a distal element effect on promoter activity in type II cells. Nevertheless, the results were qualitatively similar between the two cell types for deletions across the proximal element.

During the revision of this manuscript, Ramirez and colleagues (45) published an analysis of the T1 (=RTI40) promoter in two cell lines, one an SV40 Large T antigen immortalized neonatal rat type II cell line that expresses endogenous RTI40 (SV40TII) and the other a human lung fibroblast cell line that does not express RTI40 (IMR90). Our results with the nested 5'-deletion constructs in H441 cells (Figure 4) are similar to those of Ramirez and colleagues in SV40TII cells. In particular, we are in basic agreement on the boundaries of the proximal and distal elements. Our transfection studies include the internal deletion of the distal element (Figure 4, bottom) and the definition of both of these elements as enhancers (Figure 6). Ramirez and coworkers also proposed that the proximal region may control, at least in part, cell specific expression of the T1 promoter. Their data indicate that the proximal element contains functional binding sites for thyroid transcription factor I (at 120), SpI (at 94), and hepatic nuclear factor 3 (at 106) families of proteins. We have identified a distinct TTF1 binding site homology at position 40, immediately adjacent to the TATA box. Unfortunately, our current data cannot address cell specificity of RTI40 expression primarily because RTI40 promoter reporter genes are also highly active in all nonpulmonary cell lines we have examined to date (Rat1 fibroblasts, mouse F9 embryocarcinoma, and COS7 monkey kidney) (J. N. Vanderbilt, in preparation). In particular, the activity of the proximal element appears similar between the Rat1 fibroblasts and H441 cells. Thus, although Ramirez and coworkers observed a difference in proximal element phenotype between a rat type II cell line and a human lung fibroblast line, this difference does not hold for all fibroblast lines including those of rat origin. We believe that a more complete understanding of the determinants of cell specificity will require improvements in type II cell transfection techniques and/or a transgenic approach.

In summary, we have characterized the gene and promoter for RTI40, a cell surface antigen expressed on alveolar type I epithelial cells. This large gene spans 35 kb and contains six exons, five introns, and at least six ID repetitive elements. Mutagenesis of the 5' flanking DNA led to the identification of two elements, one distal to the transcription start site between 1247 and 795 and the other proximal between 163 and 81, that contribute to promoter activity in transfected H441 cells. The promoter was also active in transfected type II cells; in this case, deletion across the proximal region markedly reduced promoter activity, whereas deletion across the distal region was without effect. Experiments are in progress to investigate the cis- and trans-acting factors that contribute to the transcriptional activity of these control regions. In this way, we hope to establish whether or not the regulatory mechanisms responsible for cell- and tissue-specific expression of RTI40 also act on other genes that alter the differentiated state of the alveolar epithelium.