Supporting InformationGeng et al. 10.1073/pnas.1007293108SI Experimental ProceduresGeneration of Fstl1−/− Mice. Fstl1flox/+ mice (129S4 × C57BL/6J)were generated by standard homologous recombination at ModelAnimal Research Center, Nanjing University, China. In thesemice, Fstl1 exon 2 encoding the signal peptide was flanked by loxPsequences. Fstl1flox/+ mice were then mated to EIIa-Cre trans-genicmice (FVB/N; Jackson Laboratory), in which the adenovirusEIIa promoter directs expression of the Cre enzyme in earlymouse embryos (two- to eight-cell stage) to achieve homologousrecombination between LoxP sites and triggers the deletion of theexon 2 in all cells of the developing animal, including the germcells that transmit the genetic alteration to progeny. Deletion ofthe exon 2 results in loss of its signal peptide and disrupts its ORF,leading to loss of Fstl1 expression. Because of themosaic activitiesof Cre recombinase, the first generation of EIIa-Cre;Fstl1flox/+

might be chimeric with different deletions. Therefore, the chi-meric offspring were backcrossed with C57BL/6J to generateFstl1+/− mice (129S4 × C57BL/6J × FVB/N), which then wereintercrossed for production of Fstl1 deficient (Fstl1−/−) mice.Southern blotting analysis of progenies from chimeric micedemonstrated the expected gene replacement in Fstl1 locus andgerm line transmission of the targeted allele. Their WT litter-mates were used as control. All animal experiments were ap-proved by the Institutional Animal Care and Use Committee ofthe Nanjing University Model Animal Research Center.

Genotyping and Southern Analysis. Mouse genotyping was per-formed on genomic DNA isolated from mouse tails by PCR withthe following primer sets Ff1, 5′-TCC CAC CTT CGC CTC TAACT-3′; and Rf4, 5′-GAA CTC TGC GGC TGC TCT G-3′. A560-bp or an approximately 350-bp fragment was produced fromthe WT allele or the null allele, respectively. Genomic Southernblots of mice tail biopsies were performed using standard pro-cedures. A 500-bp fragment, located at the 5′ terminus of the 5′homologous arms, was used as a probe which was obtained byPCR with the primer set (5′-AAC CCT GCT TTC TGT CTG C-3′ and 5′-CAT GCC TTC CTA TTT GTT GG-3′) using genomicDNA as a template.

Western Blotting, Pull-Down Assay, and qRT-PCR Analysis. The anti-bodies were used to recognize the following proteins: Fstl1,mature-SP-C, mature-SP-B, c-Myc (A14), β-actin, and GAPDH(Santa Cruz Biotechnology); pro-SP-C and pro-SP-B (Abcam);phospho-Smad1/5, total-Smad5, total-Smad1, and phospho-Smad2 (Cell Signaling Technology); and Flag (Sigma). For pull-down assay, COS7 cells were transfected with BMPRII-flag orMyc-His-Fstl1 with Lipofectamine (Invitrogen). In vitro trans-latedMyc-His–tagged Fstl1 was concentrated by Ni-NTA-agarosebeads (Qiagen) which could capture His-tagged proteins. Afterwashing, the beads were incubated with extrogenous Flag-taggedreceptor cell lysates. The Ni-Fstl1–bounded Flag receptors werethen analyzed by immunoblotting with anti-Flag, or anti-Mycantibodies. Total RNA was isolated with TRIzol (Invitrogen),cleaned with the RNeasyMini Kit (Qiagen) and the DNA-free kit(Ambion), and then qRT-PCR was performed by using SYBRGreenER qPCR SuperMix Universal (Invitrogen) according tothe manufacturer’s protocols. Gene expression was measuredrelative to the endogenous reference gene, mouse β-actin or hu-man GUSB, using the comparative CT method described pre-viously (1). Sequences of the specific primer sets are as follows:Col2a1 (NM_031163), forward, 5′-CCA AAC CAG CCT GACAAC TT-3′; reverse, 5′-TCT AGC ATG CTC CAC CAC TG-3′;

Sftpc (NM_011359), forward, 5′-GAA GAT GGC TCC AGAGAG CAT C-3′; reverse 5′-GGA CTC GGA ACC AGT ATCATGC-3′;Cc10 (NM_011681), forward, 5′-CATGCTGTCCATCTG CTG C-3′; reverse, 5′-CTC TTG TGG GAG GGT ATC C-3′; Aqp5 (NM_009701), forward, 5′-GGT GGT CAT GAA TCGGTT CAG C-3′; reverse, 5′-GTC CTC CTC TGG CTC ATATGTG-3′; T1α (NM_010329), forward, 5′-TGC TACTGGAGGGCTTAATGA-3′; reverse, 5′-TGCTGAGGTGGACAGTTCCT-3′; β-actin (NM_007393), forward, 5′-AGG CCA ACC GTGAAA AGA TG-3′; reverse, 5′-AGA GCA TAG CCC TCG TAGATG G-3′; SFTPC (NM_003018), forward, 5′-TCT CCA CATGAG CCA GAA ACA C-3′; reverse, 5′-CGT TGC TGG GCTTCCG-3′; andGUSB (NM_000181), forward, 5′-CTCATTTGGAAT TTT GCC GAT T-3′; reverse, 5′-CCG AGT GAA GATCCC CTT TTT A-3′.

Morphological Analysis. Whole embryos or lungs were fixed in 4%paraformaldehyde in PBS solution at 4 °C and embedded inparaffin. Sections (5 μm) were mounted on slides and stained withH&E. Lung sections were also stained with PAS. Antibodies usedfor IHC were pro-SPC and Col2a1 (Abcam); TTF1, CC10, andT1a (Santa Cruz Biotechnology); and p-HH3 (Cell SignalingTechnology). Trachea of the E18.5 pups were removed understereoscope and stained with Alcian blue. Transmission EM wasperformed with lung tissue obtained from E18.5 Fstl1−/−mice andlittermate controls after fixation in glutaraldehyde according tothe standard procedures. Cell proliferation assay was performedby counting p-HH3–positive cells, as well as the total cell numberin three adjacent slides from three different WT or Fstl1−/− em-bryos. Cellular proliferation was also measured by nuclear in-corporation of BrdU in the lungs from E15.5 and E18.5 embryos.For BrdU labeling, 0.1 mg/g body mass of BrdU (Roche) wasinjected intraperitoneally into pregnant female mice. Injectedanimals were killed 1 h later (E15.5) or 2 h later (E18.5), andembryos were then processed for paraffin sections. Five-micrometer sections were treated with 2 NHCl for 1 h, neutralizedwith PBS solution six times for 10 min each, and immunostainedovernight at 4 °C with a mouse monoclonal antibody to BrdU(1:50 dilution; Roche). Samples were incubated with HRPpolymer–goat anti-mouse IgG (Maixin_Bio) for 15 min. Signalsdetected by using 3,3-diaminobenzidine substrate (Maixin_Bio).Slides were slightly counterstained with hematoxylin (Sigma) andexamined by light microscopy. Quantitative measurement wasmade with Image Pro Plus 6.0 software (Media Cybernetics). Thelung sections in each group were coded, and representative im-ages were acquired with a Leica DM3000 microscope by an in-vestigator who was blind to the identity of the slides. Ten randomfields at a magnification of 400× were selected from each sample(n = 4 for each group). The images were taken under the sameexposure setting.

Cell Culture and Saccular Explant Culture. Human pulmonary epi-thelial cells A549 and hepatoma epithelial cells Hep3B wereobtained from American Type Culture Collection. ChondrogenicATDC5 cells were obtained from the Riken Cell Bank. Cells weremaintained in DMEM/F12 (Gibco) supplemented with 5% FBS(HyClone) and antibiotics in 5% CO2 at 37 °C in a humidifiedatmosphere. ATDC5 cells started to differentiate when replacingwith differentiation medium containing 2% ITS media supple-ment (Sigma). At 48 h after transfection with Fstl1-expressingplasmid or pcDNA3.1, stable transfectants were selected withG418 (Sigma) for 14 d. Individual clones were picked and am-

Geng et al. www.pnas.org/cgi/content/short/1007293108 1 of 4

plified and cells were collected for analysis of Fstl1 levels. Cel-lular proliferation was measured by the MTT assay manufac-tured by Amresco. Saccular explants culture was performed asdescribed (2). Briefly, E15.5 fetal mouse lungs were isolated anddissected free of surrounding structures. The lung tissues wereminced into 0.5- to 1-mm3 cubes and cultured on an air–liquidinterface using permeable supports (Transwell, 0.4-μm pore size;Corning) and serum-free DMEM. Explants were cultured at 37 °Cin 95% air/5% CO2 for up to 72 h.

Luciferase Assay. Hep3B cells were cotransfected with luciferasereporters and indicated plasmids [pc-Fst was provided byZhengpin Xu (Zhejiang Unversity, Hangzhou, China) and HenryT. Keutmann (Massachusetts General Hospital, Boston, MA)], aswell as Renilla-luciferase (20 ng) as an internal control, by usingLipofectamine (Roche). One day after transfection, they weretreated with BMP4 (20 ng/mL unless otherwise indicated) inMEM containing 0.2% FBS. After washing with PBS solution,cells were harvested and the luciferase activity of cell lysates was

determined by using a luciferase assay system (Promega) as de-scribed by the manufacturer. Total light emission during theinitial 20 s of the reaction was measured in a luminometer (LumatLB 9501; Berthold). All assays were repeated in triplicate.

SPR Analysis. Real-time binding experiments were performed onthe Biacore 2000 apparatus (Biacore). Purified Fstl1 was immo-bilized on the sensor-chip surface (CM5, certified grade; Biacore).BMP4, TGF-β1, activin A, NGF, and EGF ligand proteins (R&DSystems) ran over Fstl1 sensor chip. After ligand flow ended,dissociation was monitored by a decrease in the resonance units.Kinetic experiments were performed using ligand concentrationsranging between 6.5 nM and 210 nM. Binding affinity is expressedby Kd, and a higher Kd means lower affinity of binding.

Statistical Analysis. Data are expressed as mean ± SEM. Differ-ences in measured variables between experimental and controlgroup were assessed by using two-sided Student t tests. Resultswere considered statistically significant at P < 0.05.

1. Ning W, et al. (2004) Comprehensive gene expression profiles reveal pathways relatedto the pathogenesis of chronic obstructive pulmonary disease. Proc Natl Acad Sci USA101:14895–14900.

2. Prince LS, Okoh VO, Moninger TO, Matalon S (2004) Lipopolysaccharide increasesalveolar type II cell number in fetal mouse lungs through Toll-like receptor 4 and NF-kappaB. Am J Physiol Lung Cell Mol Physiol 287:L999–L1006.

Fig. S1. Schematic representation of the gene knockout strategy. (A) Fstl1flox/+ was crossed with EIIa-Cre transgenic mice, resulting in exon 2 deletion. (B)Southern blot analysis of progenies from chimeric mice and C57BL/6J, indicating the WT allele (6 kb) and loxP mutated allele (4.2 kb), resulting from an EcoRIrestriction enzyme digest.

Geng et al. www.pnas.org/cgi/content/short/1007293108 2 of 4

Fig. S2. Tracheal malformations in Fstl1−/− mice. (A) Transverse sections of the E18.5 trachea of a WT and two homozygous mice before (E18.5, Left) and afterbreath (P0, Right) showed an increase in the lumen diameter (asterisk), dispersion and discontinuity of cartilage ring (arrow), and disorganization of theepithelial layer in mutants. (Scale bars, 100 μm.) (B) Transverse sections of the middle portion of the trachea of a WT mouse and two homozygous mice. AtE13.5, mutant trachea showed a slight enlargement of the lumen. At E15.5, the tracheal malformations were obvious in mutant sections. (Scale bars, 100 μm.)(C) The disorganized tracheal epithelial layer of Fstl1−/− mice at E18.5 contained bigger cells with shorter and disordered cilia, whereas columnar andstraightforward ciliated pseudostratified epithelial cells were observed in WT tracheas. (Scale bars, 100 μm.)

Fig. S3. Abnormal atelectasis and hypercellularity phenotype in Fstl1−/− mice. (A) Lung sections were prepared from Fstl1−/− mice at the indicated gestationalstages and stained with H&E. No significant differences were detected between the WT and mutant embryos at E15.5 and E16.5, but dramatic differences werevisible from E17.5 onward. A condensed appearance was observed in Fstl1−/− lungs during this period. (Scale bars, 200 μm.) (B) Cell proliferation by p-HH3 IHC.Immunostaining of p-HH3 in WT and Fstl1−/− lungs at E15.5 and E18.5, respectively. (Scale bars, 50 μm.) (C) Cell proliferation was measured by BrdU labeling forlungs at E15.5 and E18.5. (Scale bars, 150 μm.) (D) Quantification of cell proliferation by BrdU labeling in E15.5 WT and Fstl1−/− lungs. The graph represents themean of four independent experiments showing the comparison of the percentages of BrdU-positive cells in the epithelium and mesenchyme between WT andFstl1−/− lungs. (E) PAS staining showed dispersed cytoplasmic glycogen (pink stain) in Fstl1−/− mutant lungs. (Scale bars, 50 μm.)

Geng et al. www.pnas.org/cgi/content/short/1007293108 3 of 4

Fig. S4. Fstl1 modulated BMP4/Smad1/5/8 signaling via binding to BMP4. (A) Phosphorylated Smad2 in lung tissues from WT and Fstl1−/− embryos at E18.5. (B)Fstl1 inhibited BMP4-induced expression of the reporters GCCG-luciferase activities in Hep3B cells. The construct of GCCG-luciferase reporter (0.5 μg) wascotransfected with Fstl1 construct (pc-Fstl1; +, 0.25 μg; ++, 0.50 μg) in Hep3B cells as indicated. The transfected cells were treated with BMP4 (20 ng/mL) for 16 hand harvested for luciferase assay. All of the experiments for luciferase assay were performed by cotransfection of Renilla (20 ng) as an internal control. (C)Fstl1 had no effect on constitutively activated ALK6 or ALK1. The construct GCCG-luciferase reporter (0.5 μg) was cotransfected with constructs of caALK6 (0.1μg), caALK1 (0.1 μg), and pc-Fstl1 (0.3 μg) in Hep3B cells as indicated. 36 h after transfection, cells were lysed for luciferase assay. Each experiment was per-formed in triplicate, and the data represent the mean ± SEM of three independent experiments after normalized to Renilla activity. (D) Fstl1 pull-down assay.Fstl1 did not bind to type I receptor (ALK6; Right) of BMP4. In vitro translated Myc-His–tagged Fstl1 was concentrated by Ni-NTA-agarose beads, which couldcapture His-tagged proteins. After washing, the beads were incubated with extrogenous Flag-tagged receptor cell lysates. The Ni-Fstl1 bounded Flag receptorswere then immunoblotted with anti-Flag antibody (Upper). The blot was further developed with anti-Myc antibody to confirm the presence of Fstl1 (Lower).(E) BMP4 competed with Fstl1 in binding BMPRII. Increased concentration of BMP4 was added to cell lysates to compete for Fstl1 binding.

Fig. S5. Reducing BMP4 signaling activity rescued atelectasis phenotype of Fstl1−/− embryonic lung explants. E15.5 lung explants from WT and Fstl1−/− lit-termates were cultured for 48 to 54 h in the absence (CTL) or presence of Noggin (500 ng/mL). (A) Bright-light microscopy images of sacculars at the explantperiphery showing saccular dilation in Noggin-treated Fstl1−/− explants. (Scale bars, 200 μm.) (B) Quantification of saccular area at the periphery in explantsfrom each experimental group. The graph represents the mean of four independent experiments. (C) Noggin inhibited the increased Sftpc mRNA expressionlevels in Fstl1−/− fetal lung explants. The data represent the mean ± SEM of three independent experiments.

Geng et al. www.pnas.org/cgi/content/short/1007293108 4 of 4