Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Pulmonary surfactant is a developmentally and hormonally regulated lipoprotein complex synthesized and secreted by alveolar type II cells. Surfactant is essential for normal lung function in that it prevents alveolar collapse at low lung volumes (i.e., expiration). A deficiency of pulmonary surfactant in prematurely born infants leads to respiratory distress syndrome, a significant cause of neonatal morbidity.

Pulmonary surfactant contains four nonserum proteins: surfactant protein (SP) A, B, C, and D. SP-A, the most abundant surfactant protein, is a highly modified 30 to 35 kD sialyglycoprotein (1, 2) that contains both collagenous (3) and carbohydrate binding domains (4, 5). The presence of these domains places SP-A in the family of collagenous C-type lectins along with SP-D (6) and mannose binding protein (MBP) (7). SP-A has been shown to be involved in the metabolism (8), structure (9), and function (10) of pulmonary surfactant, whereas the role of SP-D in surfactant physiology remains less defined. Both SP-A and SP-D appear to participate in local host defense and inflammatory processes of the lung (11), which is in keeping with their classification as C-type lectins (12).

The human SP-A locus consists of two very similar but nonidentical genes, SP-A1 (13) and SP-A2 (14), each approximately 5 kb in length. This is in contrast to mice (15), rats (16), rabbits (17), and dogs (18), where SP-A exists as a single-copy gene. It is thought that two SP-A1 gene products and one SP-A2 gene product associate through their collagenous domains to form heterotrimers (19). The human SP-A locus also contains a truncated and nonfunctional pseudogene (20) that shares 85% similarity with the SP-A 3' untranslated region (3'UTR). The human SP-A locus has been assigned to chromosome 10q22-q23 (21, 22) near the loci for SP-D (6, 22) and MBP (7). The clustering of SP-A1, SP-A2, SP-D, and MBP to the same cytogenetic region suggests that this area of chromosome 10 is involved in the evolution of collagenous C-type lectins (6, 7, 22). Recently, it has been found that SP-A1 and SP-A2 are in linkage disequilibrium, indicating close physical association (23).

The regulation of SP-A gene expression is under developmental and hormonal control and in the human appears to be quite complex (24). A number of 5' splicing variants and alleles have been described for each gene (23, 29- 31) that may explain the wide variation in SP-A mRNA levels seen among humans (32, 33). Moreover, there is evidence that the two SP-A genes are regulated differentially (34; our unpublished observations). Toward a more complete understanding of the mechanisms by which SP-A1 and SP-A2 are differentially and/or coordinately regulated, we sought to better characterize the human SP-A locus.

In this study, we used a combination of physical mapping, radiation hybrid mapping (at both medium and high resolutions), and long-range polymerase chain reaction (PCR) to determine the organization of the human SP-A locus. Radiation hybrid mapping was also used to determine the relative location of the SP-D locus and to orient both the SP-A and SP-D loci with respect to the centromere.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Genomic Clones

A bacteriophage EMBL3 human genomic library had been previously (35) constructed by subcloning partially Sau3A digested HeLa genomic DNA into the SalI site of the EMBL3 vector. This library was screened with a 750-bp BamHI/EcoRI fragment of human SP-A1 cDNA. Two positive EMBL3 clones, each with insert sizes of approximately 16 kb, were characterized in this study: H7 and H3. P1 and PAC (36) human genomic libraries were screened by Genome Systems, Inc. (St. Louis, MO) using PCR with SP-A-specific primers supplied by our laboratory. One SP-A1-positive clone (P1 267-5F) and one SP-A2-positive clone (PAC 307-L19) with insert sizes of 40 kb and 80 kb, respectively, were identified in this manner. P1 and PAC clones are designated according to clone addresses as suggested by the supplier. Primers generated from the ends of P1 267-5F were used to rescreen the PAC human genomic library (Genome Systems, Inc.) in order to perform a genomic walk from the 3' side of SP-A1. This process was repeated several times with each new genomic clone obtained from the walk. In doing so, PAC clones 144-5O, 23-I13, and 18-N24 were obtained; together these clones form a contig that spans approximately 250 to 300 kb on the 3' side of SP-A1.

Restriction Digests and Southern Hybridizations

Approximately 1 to 2 µg of cloned DNA was digested with 1 to 5 U of EcoRI, HindIII, or SacI (New England Biolabs, Beverly, MA). Digestion products were subjected to electrophoresis through a 1% agarose gel and transferred under vacuum (Pharmacia LKB, Uppsala, Sweden) onto Genescreen (NEN, Boston, MA) nylon membrane in 1 N NaOH after depurination in 0.1 N HCl. Membranes were then baked at 80°C under vacuum. Southern hybridizations were performed using probes that were either PCR products specific for the SP-A 3'UTR or purified restriction fragments. Probes were radiolabeled with [³²P]dCTP using random primer labeling (Prime-it II kit; Stratagene, La Jolla, CA). Approximately 5 × 10⁵ to 1 × 10⁶ cpm of probe per milliliter of hybridization solution were used in Southern hybridizations. Prehybridization was performed at 42°C in 1 ml/10 cm² of 90% (vol/vol) Hybrisol I (Oncor, Gaithersburg, MD), 5× Denhardt's solution (0.01% each of bovine serum albumin [BSA], Ficoll, and polyvinylpyrrolidone), and 2 µg/ml denatured fish sperm DNA. Probe was added in combination with 2 µg/ml denatured fish sperm DNA to the prehybridization solution. After 14 to 18 h of hybridization at 42°C, the membranes were washed twice in 1× SSC (0.15 mM NaCl, 0.015 mM sodium citrate [pH 7.5])/0.5% sodium dodecyl sulfate (SDS) at 25°C for 20 min and then once each in 0.5× SSC/0.5% SDS at 50°C and then at 65°C for 20 min. Membranes were then exposed to X-ray film (Kodak, Rochester, NY).

Sequencing of Genomic Clones

In order to confirm the identity of sequences within our genomic clones, to develop long-range PCR primers, and to generate primers used in amplifying sequence tagged sites (STSs) for radiation hybrid mapping (see the following), the EMBL3, P1, and PAC clones were partially sequenced. Sequencing primers were derived from the published sequences for SP-A1 (13), SP-A2 (14), and the SP-A pseudogene (20) or from vector sequences. Primers MBL (5'-GCTTATCTGCTTCTCATAGAGTCTTGC-3') and MBR (5'-ATAACGATCATATACATGGTTCTCTCC-3') are specific for the left and right arms of the EMBL3 vector, respectively, and were used to sequence the insert ends of clones H3 and H7. T7 and SP6 primers were used in the sequencing of insert ends of the P1 and PAC clones. Sequencing of the EMBL3 genomic clones was performed using the f-mol cycle sequencing kit (Promega, Madison, WI) and consisted of 35 cycles with denaturation at 94°C for 30 s, annealing at 42°C for 30 s, and extension at 72°C for 1 min. P1 and PAC DNA was precipitated with polyethylene glycol in 1.5 M NaCl and resuspended in H₂O before sequencing. Sequencing reactions of the P1 and PAC clones consisted of 35 cycles of denaturation at 94°C for 30 s followed by annealing at 50°C for 30 s and extension at 72°C for 1 min. All sequencing reactions were run on 6% PAGE/7 M urea gels that were dried under vacuum and exposed to X-ray film (Kodak). Primers to be used in long-range PCR and radiation hybrid mapping were designed using the software OLIGO v4.0 (National Biosciences Inc., Plymouth, MN).

Long-range PCR

Primers used in long-range PCR are described in Table 1. The protocol for long-range PCR was modified from that previously described (37). Briefly, the reaction mixture contained 20 mM Tris-HCl (pH 8.6), 150 µg/ml BSA, 16 mM (NH₄)₂SO₄, 3.5 mM MgCl₂, 250 µM of each dNTP, 30 mM Tris base, 8% glycerol, 100 to 200 ng of template DNA, and 100 ng of each primer. After an initial denaturation of 1 min, 2.5 U of AmpliTaq (Applied Biosystems, Foster City, CA) combined with 0.15 U of Pfu DNA polymerase (Stratagene) were added. All long-range PCRs were performed using thin-walled tubes (USA/Scientific Plastics, Ocala, FL) in a total volume of 100 µl. Cycle conditions depended on the primers used, but typically reactions consisted of 30 to 35 cycles of an 8-s denaturation at 94°C, a 30-s annealing at 58 to 62°C, and a 4- to 16-min extension at 68°C.

Radiation Hybrid Mapping

Radiation hybrid (RH) mapping is a technique used to construct high-resolution genomic maps and is described in detail by Cox and colleagues (38). RH mapping was performed using the whole-genome Stanford TNG3 (high resolution) and G3 (medium resolution) radiation hybrid panels, both of which were purchased from Research Genetics (Huntsville, AL). A number of STSs were tested against the 90 human/hamster hybrid DNAs of the TNG3 panel and the 83 human/hamster hybrid DNAs of the G3 panel. PCR amplifications were carried out with sets of primers that amplify STSs identified by sequencing the ends of our genomic clones and/or were known to span the SP-A and SP-D loci. Information on the STSs amplified in radiation hybrid mapping is shown in Table 2. We also used two sets of primers that amplify expressed sequence tags (ESTs), WI-6610 and WI-11484, that have been placed on the Whitehead Institute/MIT Center for Genome Reseach (WICGR) low-resolution (Genebridge 4 panel) radiation hybrid map. Mapping data for these ESTs are available from WICGR via the World Wide Web at http://www-genome.wi.mit.edu. PCRs were performed in 96-well plates using a TwinBlock easycycler (ERICOMP, Inc., San Diego, CA) in 25 µl of 1.5 mM MgCl₂, 0.05 mM of each dNTP, and 1× PCR buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl), 50 ng of each primer, and 2.5 U of AmpliTaq DNA polymerase (Applied Biosystems). Cycle conditions depended on the primers used, but generally 30 to 35 cycles of a 30-s denaturation at 94°C, a 30-s annealing at 56 to 60°C, and a 30-s extension at 72°C were used. PCRs were performed in duplicate. The scoring of clones as positive, negative, or ambiguous for an STS was accomplished by visualization of PCR products on ethidium-stained agarose gels. Two-point and multipoint analyses of the radiation hybrid data were performed using the statistical software package RHMAP v3.0 (39), obtained via the World Wide Web at http://www.sph.umich.edu/group/statgen/software. This software package consists of three statistical programs: RH2PT, RHMAXLIK, and RHMINBRK.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Overall organization of the human SP-A and SP-D loci at three different levels of resolution as determined by radiation hybrid mapping. The low-resolution map was constructed by the Whitehead Institute/MIT Center for Genome Research. The medium- and high-resolution maps are from this study. The positions of STSs that could be ordered at an odds ratio greater than 1,000:1 (LOD score > 3) are shown. Distances between markers, expressed in centirays, are shown to the left of each radiation hybrid map and distances in kilobases (when known) are shown to the right. For the low-resolution map, reference genetic markers are shown with their corresponding positions in centimorgans (cM) from the top of chromosome 10 linkage map in parentheses.  indicates that no evidence for linkage at an odds ratio greater than 1,000:1 (LOD score > 3) was found between SP-A1 and 530/531.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Analysis of Two Overlapping Genomic Clones Reveals Close Proximity of the SP-A Pseudogene to SP-A1

Initial characterization of the EMBL3 genomic clones revealed that H7 contained the entire SP-A pseudogene and about 14 kb of flanking sequences. H3 was found to contain only the 5' half of SP-A1 (approximately bp 1-1800) in addition to 13 kb of 5' flanking sequence. Among the fragments produced by digestion of H3 and H7 with EcoRI are two that are common to both clones: a 1.2-kb and a 7.4-kb fragment (Figure 1A, lanes 2 and 3). Likewise, digestion with HindIII yielded 1.8-kb, 2.2-kb, and 3.5-kb fragments from both clones (not shown). These fragments do not come from the EMBL3 vector; thus, it appeared that there was an overlap between H7 and H3. In order to verify this overlap, the 1.2-kb EcoRI fragment from H3 was isolated, radiolabeled, and then used as a probe in Southern hybridizations. Figure 1B shows that the 1.2-kb EcoRI fragment from H3 hybridizes to the same size EcoRI fragment of H7 (lane 2), supporting an overlap between the clones. An overlap between H3 and H7 indicates that the SP-A pseudogene lies upstream of SP-A1 because H3 only contains the 5' half of SP-A1. An overlap between H7 and the clone P1 267-5F (which contains the entire SP-A1 gene) was also found, supporting close physical proximity of the SP-A pseudogene and SP-A1.

View larger version (43K): [in this window] [in a new window]   Figure 1.   Restriction digests and Southern hybridizations of EMBL3 clones H3 and H7 show an overlap and reveal the orientation of the SP-A pseudogene relative to SP-A1. (A) H7 DNA was digested with either SacI (lane 1) or EcoRI (lane 2), and H3 DNA was digested with EcoRI (lane 3) and run on a 1% agarose gel. M = 1-kb ladder. Arrows point to the 7.4-kb and 1.2-kb EcoRI fragments common to both genomic inserts. (B) The gel from A was transferred to a Genescreen nylon membrane and probed with the radiolabeled 1.2-kb EcoRI restriction fragment derived from H3. Top arrow points to the 30-kb SacI fragment of H7 that contains the overlap and thus the SP-A pseudogene/SP-A1 intergenic region. Bottom arrow points to the 1.2-kb EcoRI restriction fragment common to H7 and H3. (C) The same membrane as in B was stripped and rehybridized to a radiolabeled PCR product derived from the 3' end of the SP-A 3'UTR (common to all three SP-A genomic sequences), which is distal to the pseudogene SacI site. Arrow points to the 15-kb SacI fragment of H7 that contains the majority of the SP-A pseudogene.

Orientation of the SP-A Pseudogene Relative to SP-A1

The enzyme SacI cuts within the SP-A pseudogene once at bp 646 (20) and digestion of clone H7 with SacI yields two fragments with approximate sizes of 30 kb and 15 kb (Figure 1A, lane 1). SacI does not cut within the EMBL3 left or right vector arms, which are 20 kb and 9 kb in length, respectively. Thus, the overlap between H7 and H3 must reside within the 30-kb SacI fragment (because this is the only SacI fragment large enough to contain the 1.2-kb and 7.4-kb EcoRI fragments as well as an EMBL3 vector arm). This is confirmed by the hybridization of the 1.2-kb EcoRI fragment from H3 to the 30-kb SacI fragment of H7 (Figure 1B, lane 1). A PCR-generated SP-A 3'UTR fragment derived from sequences downstream of the pseudogene SacI site served as probe in the Southern hybridization shown in Figure 1C. The SP-A 3'UTR probe is shown to hybridize to a 2.4-kb EcoRI fragment in Figure 1C (lane 2); this fragment corresponds to the majority of the SP-A pseudogene in addition to 3' flanking sequences. There is no hybridization of this probe to the H3 digests in Figure 1C (lane 3) because this clone only contains the 5' half of SP-A1 (no 3'UTR sequences). The 3'UTR probe hybridizes to the 15-kb SacI fragment of H7 (Figure 1C, lane 1) and not to the 30-kb SacI fragment, indicating that the SP-A pseudogene is in a reverse orientation relative to SP-A1. A partial restriction map of the SP-A pseudogene and SP-A1 intergenic region (IGR) is shown in Figure 2.

View larger version (9K): [in this window] [in a new window]   Figure 2.   Partial restriction map of the SP-A pseudogene/SP-A1 intergenic region. Boxes indicate the SP-A pseudogene (short box) and SP-A1 (longer box). Shaded areas show the area of sequence similarity between the two. Dark bar represents flanking and/or intergenic sequences. EcoRI sites are shown as vertical lines labeled "R." The SacI site found within the SP-A pseudogene is designated with a vertical line labeled "S." Arrows above the restriction map indicate relative positions and orientations of primers used in long-range PCR. Thin open bars below the restriction map represent the overlapping genomic clones that span the area between the SP-A pseudogene and SP-A1.

Long-range PCR Amplification of the SP-A Pseudogene/SP-A1 IGR

In order to verify the results obtained from the analysis of the EMBL3 clones, we attempted to amplify the IGR between the SP-A pseudogene and SP-A1 from genomic DNA using long-range PCR. Primers (Figure 2 and Table ) were generated as described in MATERIALS AND METHODS. Long-range PCR using the primer pair 413/291 resulted in a 3.2-kb band from both genomic DNA (Figure 3A, lane 1) and H7 (not shown). Long-range PCR using the primer pair 416/262 resulted in a 3.5-kb band in both genomic DNA (Figure 3A, lane 2) and clone H3 (not shown). Taken together, these PCR amplifications verify the overlap between clones H3 and H7. The entire IGR between the SP-A pseudogene and SP-A1 was then amplified using the primer pair 262/68 (Figure 3A, lane 3), which resulted in a 16-kb band, 14 kb of which comes from the SP-A pseudogene/SP-A1 IGR. We were also able to amplify the IGR using the primer pair 262/291 (not shown). These long-range PCR amplifications confirm not only the distance between the SP-A pseudogene and SP-A1, but also their relative orientation. In order for the 16-kb PCR fragment to be produced from genomic DNA, the SP-A pseudogene must be in an opposite orientation to that of SP-A1. EcoRI digestion of 262/68 long-range PCR products amplified in different reactions from two unrelated individuals is shown in Figure 3B, lanes 1 through 5. These digestions yielded the expected 1.2-kb and 7.4-kb fragments in addition to a 4.5-kb fragment from the SP-A1 end of the 262/68 PCR product, a 2.6-kb fragment from the 5' side of the SP-A pseudogene, and a 0.5-kb fragment from the pseudogene end of the 262/68 PCR product. These digestions confirm the identity of the 262/68 long-range PCR product as being the SP-A pseudogene/SP-A1 IGR.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Long-range PCR amplification of the SP-A pseudogene/ SP-A1 intergenic region. (A) Primer pairs 413/291, 416/262, and 262/68 produce 3.2-kb, 3.5-kb, and 16-kb PCR products from human genomic DNA, respectively. M1 = 1-kb marker and M2 = HindIII digest of lambda marker. (B) Lanes 1 through 5: 262/68 long-range PCR products digested with EcoRI; lane 6: EcoRI digest of clone H3. Arrows point to the 1.2-kb and 7.4-kb EcoRI fragments found within the SP-A pseudogene/SP-A1 intergenic region.

High-resolution Radiation Hybrid Mapping of the SP-A and SP-D Loci

We tested a number of STSs against the high-resolution TNG3 radiation hybrid panel. A description of these STSs is shown in Table . The mean rate of retention (i.e., the percentage of radiation hybrid clones found to be positive) for the STSs in this panel was approximately 20%. Two-point analysis of the TNG3 radiation hybrid data was performed using the program RH2PT and is presented in Table 3. Distances are represented as centirays 50,000 (cR_50,000). Using the previously established relative location and orientation of the SP-A pseudogene and SP-A1 as an anchor, the two-point analysis shows that SP-A2 is on the 5' side of SP-A1, with the SP-A pseudogene located between the two functional genes. SP-D also appears to be on the 5' side of SP-A1 and is closer to SP-A2 than to the SP-A pseudogene or SP-A1. Multipoint analysis was performed using the program RHMAXLIK and supports the two-point analysis in that the most likely order was found to be: SP-DSP-A2-pseudogeneSP-A1 (1,183 times more likely than any other order tested when using an equal retention probability model).

A direct comparison of cR_50,000 to physical distance can be made because the distance in kilobases has already been established between the SP-A pseudogene and SP-A1 (about 17 kb between the two STSs used). From the distances in centirays 50,000 shown in Table , we obtain a value of 1 cR_50,000 ~ 1.5 kb. Assuming that this correlation holds true for the entire chromosome, we can estimate the distances in kilobases between the rest of the STSs shown in Table . Using the correlation of 1 cR_50,000 ~ 1.5 kb, it appears that SP-A2 and SP-D are approximately 40 kb and 120 kb away from SP-A1, respectively. During the characterization of our genomic clones, PAC 307-L19 was identified as containing the entire SP-A2 gene, 2 kb of 5' flanking, and approximately 80 kb of 3' flanking sequences. This clone does not contain any SP-A1, SP-A pseudogene, or SP-D sequences nor does it overlap with any of the genomic clones mentioned previously. Taking these observations into account, SP-A2 must lie in a transcriptionally opposite direction than SP-A1 in order for these two genes to be no more than 40 kb apart.

Support for the relative location of SP-A2 and SP-D also comes from the contig (described in MATERIALS AND METHODS) generated by genomic walking on the 3' side of SP-A1. This PAC contig spans approximately 250 to 300 kb and contains no SP-A2 or SP-D sequences, indicating that these genes are on the 5' side of SP-A1.

Orientation of the SP-A and SP-D Loci Relative to the Centromere

We were able to orient the SP-A and SP-D loci with respect to the centromere by testing several STSs against the G3 medium-resolution hybrid panel. A description of these STSs is shown in Table . The mean rates of retention for these STSs in the G3 panel were dichotomous: WI-6610, SP-D, SP-A1, pseudogene, and SP-A2 had a retention rate of about 10%, and 530/531 and WI-11484 had a retention rate of about 24%. Results of the subsequent two-point analysis for the radiation hybrid data are presented in Table 4. Distances are represented as centirays 10,000 and data for markers found to have significant linkage (LOD score > 3) are shown. We cannot determine the relative order of the members of the SP-A locus because of the lower resolution in this radiation hybrid panel. However, the data support the relative locations of SP-A2 and SP-D as determined by the TNG3 high-resolution panel (i.e., SP-D is further away from SP-A1 than from SP-A2).

We can also conclude from the results in Table that the SP-D locus is the closest to the EST WI-6610. WI-6610 has been placed on the WICGR low-resolution (Genebridge 4) radiation hybrid map at a distance of 457.75 cR_3,000 from the top of the chromosome 10 linkage map. This EST is closer to the centromere than the SP-A locus (identified on the WICGR map as WI-7219 and located at 462.46 cR_3,000 from the top of the chromosome 10 linkage map). Thus, it appears that SP-D is closer to the centromere than the SP-A locus and the overall order is: centromere SP-DSP-A2pseudogeneSP-A1telomere. This orientation with respect to the centromere is supported by the linkage (LOD > 3) of an STS (530/531) found approximately 300 kb from the 3' side of SP-A1 and the EST WI-11484. WI-11484 has been mapped distal to SP-A at 467.61 cR_3,000 from the top of the chromosome 10 linkage group. The overall organization of the SP-A and SP-D loci at low, medium, and high resolutions is displayed in Figure 4.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Pulmonary surfactant consists of four proteins: SP-A, SP-B, SP-C, and SP-D. Both SP-A and SP-D are collagenous C-type lectins that are involved in local host defense and inflammatory processes of the lungs (reviewed in reference 11). The human SP-A locus consists of two functional genes, SP-A1 and SP-A2 (14), and a pseudogene (20). The human SP-A and SP-D loci have been localized to 10q22-q23 by both in situ hybridization (21) and somatic cell hybrid analysis (6, 22). In addition, the SP-A locus has been placed on the WICGR radiation hybrid map at 462.64 cR_3,000 from the top of the chromosome 10 linkage group (there is no information regarding the location of SP-D on this map). These mapping techniques, however, lack the resolving power to determine the exact order and distance of the members of these two loci. In the present study, we mapped the SP-A and SP-D loci at a high resolution and ordered these loci relative to the centromere. This was accomplished by using not only conventional physical mapping techniques (i.e., the analysis of overlapping clones) but newer ones such as long-range PCR and high-resolution (at around 10 kb) radiation hybrid mapping.

Radiation hybrid mapping (38) is a technique that uses X-rays of a specific dose (10,000 rads for the G3 panel and 50,000 rads for the TNG3 panel) to randomly fragment the chromosomes of a human donor cell line. The higher the dose of ionizing radiation, the smaller the fragments produced. The donor cell line is fused to a recipient cell line (generally TK CHO cells), and hybrid clones are selected and tested for the presence or absence of STSs. The closer two STSs are together, the greater the likelihood that they will be found in the same hybrid clone. Statistical analysis using the software package RHMAP v3.0 (39) can then be used to determine the best estimate of physical distance, which is expressed in centirays specific for the X-ray dose used to construct the radiation hybid panel. In this study, we used both medium- and high-resolution radiation hybrid panels (G3 and TNG3, respectively) to order the members of the SP-A and SP-D loci.

Once the relative order and orientation of the SP-A pseudogene and SP-A1 were established through physical mapping and long-range PCR, attempts were made to determine the relative location of SP-A2 by chromosomal walking. However, these attempts were undermined by the failure to detect any SP-A2 sequences 3' of SP-A1 (up to a distance of about 250 to 300 kb) and the inability to obtain true overlapping clones of H7 (on the 3' side of the SP-A pseudogene). Analysis of two SP-A-positive yeast artificial chromosomes (YACs) indicated the presence of SP-A1 and SP-A2, but not the SP-A pseudogene. This initially led us to believe that SP-A2 was on the 3' side of SP-A1 (40). However, we have subsequently concluded that either one or both YACs may contain internal deletions. Therefore, in the present study, we show the utility of high-resolution radiation mapping to determine gene order, especially when overlapping or high molecular weight clones (i.e., P1 artificial chromosomes [PACs] and YACs) are unavailable or suffer from deletions and/or chimerisms.

A direct comparison can be made between centirays 50,000 and kilobases because some of the STSs used in the high-resolution radiation hybrid mapping had also been physically mapped (namely, the SP-A pseudogene and SP-A1). We are able to estimate that 1 cR_50,000 approximates 1.5 kb. Using this correlation, we can estimate that SP-A1 and SP-A2 (which are 26.4 cR_50,000 apart; Table ) are separated by about 40 kb. Similarly, SP-A1 and SP-D are estimated to be 120 kb apart, supporting earlier findings that placed the two loci within 430 kb (22). These conclusions are based on the assumption that the correlation between centirays 50,000 and kilobases remains constant throughout the genome. In fact, when chromosome 2 STSs are tested against the TNG3 panel, we do get a similar correlation of 1 cR_50,000 ~ 1.4 kb (41). Furthermore, previous studies using lower-resolution radiation hybrid panels have shown a relatively constant relationship between centirays and physical distance (38), as in the case with the G3 panel, where 1 cR_10,000 ~ 30 kb. Two-point analysis of the TNG3 radiation hybrid data showed that the distance between SP-A1 and SP-A2 is 26.4 cR_50,000. However, when this distance is taken as the sum of the distance between SP-A2 and the pseudogene (12.6 cR_50,000) plus the distance between the pseudogene and SP-A1 (11.3 cR_50,000), it becomes 23.9 cR_50,000. Such inconsistencies have been reported before (38) and do not appear to change our general conclusions.

The orientation of SP-A2 is found to be opposite to that of SP-A1, i.e., the two genes are in a "head to head" arrangement. This conclusion is made based on the analysis of an SP-A2-positive genomic clone (PAC 307-L19) that contains the SP-A2 gene, 2 kb of 5' flanking sequence, and about 80 kb of 3' flanking sequence. This clone does not, however, contain any SP-A1, SP-A pseudogene, or SP-D sequences. In order for SP-A1 and SP-A2 to be 40 kb (26.4 cR_50,000; Table ) apart without PAC 307-L19 containing any SP-A1 sequences, the two genes must be in opposite transcriptional orientations. Of course, this conclusion is also based on the assumption that PAC 307-L19 is not chimeric. Although chimerism rates in P1 and PAC (36) clones are far lower than in YACs, the possibility still exists. A "head to head" arrangement of SP-A1 and SP-A2 raises the possibility that the two genes may share cis-acting regulatory elements. Characterization of the loci for other "head to head" genes has yielded information on shared regulatory regions; this is particularly true for the 1 and 2 chains of type IV collagen, two genes whose products, like SP-A1 and SP-A2, form heterotrimers (42).

To orient the SP-A and SP-D loci relative to the centromere, we used the medium-resolution G3 radiation hybrid panel. The combined use of medium- and high-resolution radiation hybrid mapping produces the overall map: centromereSP-DSP-A2pseudogeneSP-A1telomere. Although we are unable to determine the relative order of the members of the SP-A locus with the G3 panel alone, we can conclude that SP-D is closer to SP-A2 than it is to SP-A1. The distance in centirays between SP-D and SP-A1 is 1.4 times the distance between SP-D and SP-A2 in both the medium- and high-resolution panels. This similarity in distance proportions between the two panels supports the validity of our map. However, the physical distance between STSs suggested by using the G3 panel are greater than those of the TNG3. For example, assuming 1 cR_10,000 ~ 30 kb, SP-A1 and SP-D are about 640 kb apart and SP-A1 and the SP-A pseudogene are 200 kb apart; the latter value is incorrect as shown by physical mapping. The expansion of physical distance when using the G3 panel could be a result of the relatively low rates of retention for the SP-A and SP-D loci (compare 10% in the G3 panel to a mean of 20% in the TNG3 panel).

Although the SP-A and SP-D loci have been previously mapped close to the locus for MBP (22), we did not find significant evidence for linkage between MBP and any member of the SP-A or SP-D loci, even when using the G3 medium-resolution panel. In fact, there was no linkage found between MBP or either of the ESTs, WI-6610 and WI-11484. Analysis of the radiation hybrid data for the MBP STS using the Stanford Radiation Hybrid Mapping Server (http://shgc-www.stanford.edu) localized MBP to within 42.3 cR_10,000 (about 1,200 kb) of a genetic marker (D103567) that maps 79 centimorgans (cM) from the top of the chromosome 10 linkage group (the SP-A locus would be found between 106 cM and 119 cM; see Figure 4, low resolution). These findings suggest that the MBP and SP-D/SP-A loci are separated by a rather large distance (about 25 to 35 cM or 25,000 to 35,000 kb).

The SP-A genomic sequences seem to have arisen by a process of gene duplication. Sequence analysis of the three SP-A genomic sequences suggests that an ancestral SP-A sequence duplicated, giving rise to SP-A1 and SP-A2. Subsequent to this event, another duplication produced the SP-A pseudogene from SP-A2 (45). In this scenario, the SP-A pseudogene is closer in evolution to SP-A2 than it is to SP-A1. The fact that SP-A2 and the SP-A pseudogene are in the same relative orientation suggests that the two are closer in evolution (i.e., arisen from the same ancestral sequence) than SP-A1 and SP-A pseudogene. One can also see a parallel between physical distance and evolutionary relationships: of the C-type lectins found on chromosome 10, MBP is the farthest away from SP-A in both physical and evolutionary (22) terms. Similarly, SP-D is more related to SP-A in evolution (22) and is also physically closer to the SP-A locus than MBP. Recently, two SP-A genes, one SP-A1-like and the other SP-A2-like, have been identified in the baboon (46). The presence of two SP-A genes in the baboon indicates that the duplication of the primordial SP-A sequence into SP-A1 and SP-A2 occurred more than 26.5 million yr ago (46). It would be of interest to know whether the baboon has a SP-A pseudogene-like sequence. Absence of such a sequence would indicate that the pseudogene arose after the separation of apes from the lineage that led to humans.

Elucidation of the organization of other gene clusters has often provided information on the regulation of gene expression (42, 47). For example, the genes of the globin cluster are developmentally regulated in a manner that parallels their spatial organization (47). Thus, it is our hope that a detailed characterization of the human SP-A locus will not only give evidence of evolutionary relationships but also lead to a more complete understanding of the mechanisms involved in SP-A1 and/or SP-A2 gene expression. Given the important roles of SP-A in surfactant physiology and host defense, such an understanding could give insight into the reasons behind various clinical outcomes or lead to the design of better interventions for respiratory disease.