A Nonaggregating Surfactant Protein C Mutant Is Misdirected to Early Endosomes and Disrupts Phospholipid Recycling


Surafel Mulugeta, mulugeta@mail.med.upenn.edu


Interstitial lung disease in both children and adults has been linked to mutations in the lung-specific surfactant protein C (SFTPC) gene. Among these, the missense mutation [isoleucine to threonine at codon 73 = human surfactant protein C (hSP-CI73T)] accounts for ∼30% of all described SFTPC mutations. We reported previously that unlike the BRICHOS misfolding SFTPC mutants, expression of hSP-CI73T induces lung remodeling and alveolar lipoproteinosis without a substantial Endoplasmic Reticulum (ER) stress response or ER-mediated intrinsic apoptosis. We show here that, in contrast to its wild-type counterpart that is directly routed to lysosomal-like organelles for processing, SP-CI73T is misdirected to the plasma membrane and subsequently internalized to the endocytic pathway via early endosomes, leading to the accumulation of abnormally processed proSP-C isoforms. Functionally, cells expressing hSP-CI73T demonstrated both impaired uptake and degradation of surfactant phospholipid, thus providing a molecular mechanism for the observed lipid accumulation in patients expressing hSP-CI73T through the disruption of normal phospholipid recycling. Our data provide evidence for a novel cellular mechanism for conformational protein-associated diseases and suggest a paradigm for mistargeted proteins involved in the disruption of the endosomal/lysosomal sorting machinery.

The hydrophobic human surfactant protein C (hSP-C) secreted into the pulmonary alveolus is a product of a larger bitopic, type II integral membrane precursor protein (proSP-C) (depicted in Figure 1) that undergoes palmitoylation and a series of proteolytic cleavages to yield the 3.7-kDa mature form (1,2). Produced exclusively by alveolar type II cells, mature hSP-C is packaged into specialized lysosomal-like organelles (lamellar bodies) and released via regulated exocytosis together with other surfactant proteins and phospholipids. In the alveolus, SP-C along with another surfactant protein, SP-B, plays a critical role in the modulation of lung mechanics from its direct effects on alveolar surface tension. Through enhancement of the adsorption and spreading of surfactant phospholipids and by maintenance of interfacial film at the air–liquid interface, hSP-C promotes the efficient generation of a surfactant monolayer shown to be critical in the prevention of alveolar collapse at end expiration [for review see (1,2)].

Figure 1.

Schematic representation of the hSP-C expression constructs used in this study is shown. Additional constructs including EGFP-C1 vector and ABCA3WT•DsRed were used for negative and positive controls. Four proteolytic cleavages of SP-C propeptide at sites “a to d” occur in alveolar type II cells. In contrast, A549 cells cleave the propeptide only once at site “a”. The cytosolic (NH2) and luminal (COOH) domains are indicated.

Interstitial lung disease (ILD) represents a diverse group of disorders of largely unknown etiology characterized by variable types and degree of interstitial and alveolar inflammation, fibrosis, and impaired gas exchange. Mutations in genes critical for surfactant metabolism and function are now recognized as one of the etiological causes of ILD (3). Of these, the gene encoding hSP-C (SFTPC) has been seminal in establishing this linkage with multiple reports of both sporadic cases as well as an autosomal dominant inheritance pattern with variable penetrance. Histological phenotypes of mutation-associated SFTPC ILD range from chronic pneumonitis of infancy and nonspecific interstitial pneumonitis (NSIP) in children to usual interstitial pneumonia (UIP) and idiopathic pulmonary fibrosis (IPF) in adults (4,5). The great majority of SFTPC mutations are located in the distal C-terminus of proSP-C termed the BRICHOS domain (6), a region of ∼100 amino acids with sequence and structure homology of proteins most notably found in association with familial neurodegenerative dementias. Using a variety of in vitro models, expression of the BRICHOS mutant hSP-C cDNAs results in intracellular aggregation of misfolded protein, ER stress and subsequent induction of apoptotic cell death (7–10).

In addition to these aggregation-prone hSP-C mutations, a different phenotype is emerging for a second group of missense mutations situated outside the BRICHOS domain, but within the C-terminal region of the propeptide adjacent to the transmembrane domain (6,11). Among these, a heterozygous substitution mutation of threonine for isoleucine at amino acid 73 (hSP-CI73T) of the proprotein accounts for ∼30% (over 40 cases reported) of all hSP-C mutations associated with pulmonary disease (11–17) in both de novo (sporadic) as well as in inherited (autosomal dominant) cases. While symptomatic patients with an I73T mutation show some variability in clinical phenotype and course most likely due to the effects of additional modifier genes and/or environmental interactions (4,5,11,17,18), the associated histological and biochemical abnormalities most often include a pattern of NSIP with alveolar type II cell hyperplasia, cellular thickening of alveolar septa and preserved alveolar architecture, often accompanied by evidence of alveolar accumulation of phospholipids and protein (lipoproteinosis). Analyses of the broncho-alveolar lavage of patients with hSP-CI73T also show accumulation of abnormal C-terminal proSP-C products (11,19,20) although the consequences of their presence toward surfactant function is unknown. Moreover, in contrast to hSP-C BRICHOS mutations, in vitro expression of hSP-CI73T fails to either induce ER stress responses, or activate caspase 4 mediated apoptosis (9) suggestive of a mechanism for cellular dysfunction distinct from those elicited by grossly misfolded BRICHOS proteins.

Given the central role of the alveolar epithelium in surfactant metabolism and the unique consequences for the lung resulting from expression of the hSP-CI73T mutation including disrupted surfactant homeostasis, we hypothesized that the trafficking and processing as well as the functional consequences of hSP-CI73T expression in vitro would be distinct from SP-C BRICHOS mutations. In this study, we demonstrate that unlike either its wild-type counterpart or SP-C BRICHOS mutants, the hSP-CI73T proprotein is initially mistargeted to the plasma membrane and subsequently internalized to early endosomes accompanied by aberrant post-translational processing. Functional assays using H3-labeled dipalmitoylphosphatidylcholine (DPPC), the major surfactant lipid species, demonstrated significant reduction of both lipid uptake and degradation by cells stably expressing hSP-CI73T. Thus, expression of this non-BRICHOS mutant results in a unique cellular phenotype distinct from previously characterized SFTPC mutations.


The human I73T mutant propeptide is abnormally trafficked to early endosome compartments

To characterize the influence of the I73T missense mutation on the biosynthesis of proSP-C, as well as on cellular homeostasis, a previously described human proSP-C cDNA clone was systematically altered where isoleucine was substituted with threonine at residue 73 by site-directed mutagenesis and fused to the Enhanced Green Fluorescent Protein (EGFP) reporter (Figure 1) (10,21–23). Because proSP-C is normally trafficked by homomeric association during its biosynthesis (24), initial studies were performed in transfected cell lines lacking endogenous SP-C to minimize interference from oligomeric association with an endogenous wild type isoform.

When wild-type (EGFP•hSP-CWT) and mutant hSP-C (EGFP•hSP-CI73T) isoforms were transiently transfected into A549 cells (steady-state expression, as determined by immunofluorescence microscopy), the proteins segregated into two distinct patterns of subcellular localization. The EGFP•hSP-CWT fusion was predominantly localized in cytosolic vesicles which were also positive for CD63 (a marker antigen for type II cell lamellar bodies and lysosomes) and negative for EEA1 (an early endosome marker) (Figure 2A) as previously described (10,24). In contrast, the EGFP•hSP-CI73T mutant isoform was preferentially trafficked to compartments positive for EEA1 and predominantly negative for CD63 (Figure 2B). In addition, a significant pool of EGFP•hSP-CI73T was also observed on the plasma membrane (Figure 2B, arrows), suggesting discrete routing differences between the two hSP-C isoforms.

Figure 2.

EGFP•hSP-CI73T is trafficked to early endosome vesicles. A and B) Representative epifluorescent images of A549 cells 24 h following plasmid introduction of EGFP•hSP-CWT (A) and EGFP•hSP-CI73T (B) immunostained with Texas Red-conjugated CD63 and EEA1 antibodies, respectively. Solid boxes from the merged images are magnified to illustrate colocalization or noncolocalization (right panels). Bar, 5 µm. C and D) Analysis of the subcellular localization of EGFP•hSP-C isoforms by density gradient fractionation. After 24 h of EGFP•hSP-CWT and EGFP•hSP-CI73Ttransfection, A549 cells were harvested and normalized for equal protein concentration per sample. Post-nuclear supernatants were subsequently prepared and loaded on top of discontinuous sucrose gradient as described in Materials and Methods. Representative immunoblotting for EGFP (C and D, top panels), LAMP 1 and EEA1 (C and D, bottom two panels) of fractionated samples and corresponding histograms of quantified bands are from at least three separate experiments. *p ≤ 0.005.

The intracellular distribution of the hSP-C isoforms in subcellular fractions was further characterized biochemically using density gradient centrifugation and immunoblotting. Lamp1, another lamellar body/lysosome marker, overlapped with EGFP•hSP-CWT in higher density sucrose fractions (Figure 2C), while EGFP•hSP-CI73T was detected in lower density fractions which contained EEA1 (Figure 2D).

Temporal expression patterns for hSP-CI73T are distinct from those of hSP-CWT

To further investigate SP-C expression patterns, co-localization studies were performed using EGFP•hSP-CWT or EGFP•hSP-CI73T and a DsRed fusion construct containing the human ATP binding cassette class A3 (hABCA3), a polytopic transmembrane lipid transporter protein that has been shown to be abundant in the limiting membrane of lamellar bodies of alveolar type II cells (25,26) and lysosomes of transfected cell lines (27). Using lipofectamine to accelerate early expression, EGFP•hSP-CWT or EGFP•hSP-CI73T and DsRed-tagged wild-type hABCA3 (hABCA3WT•DsRed) were co-expressed in A549 cells. At 12–16 h following transfection, co-expression of EGFP•hSP-CI73T with hABCA3WT•DsRed was virtually undetectable (Figure 3B (top row),D). However, 24–36 h following transfection, increasing amounts of vesicle-colocalization (∼45%) of EGFP•hSP-CI73T and ABCA3 occurred (Figure 3B (bottom row),D) and was similar to CD63 colocalization (Figure 3C). In contrast, co-localization of EGFP•hSP-CWT with hABCA3WT•DsRed was nearly 100% at all time points examined (Figure 3A) with quantitative counting (Figure 3D) suggesting that the biosynthetic pathway taken by the two SFTPC isoforms was divergent.

Figure 3.

Time-dependent trafficking of EGFP•hSP-CI73T. Representative epifluorescent images of A549 cells at different time intervals following concomitant expression of ABCA3WT•DsRed with EGFP•hSP-CWT (A) or EGFP•hSP-CI73T (B) in cells transiently transfected using Lipofectamine transfection method. Times shown are 12–16 h and 24–36 h post-transfection. Bar, 5 µm. C) Partial co-localization of EGFP•hSP-CI73T in CD63+ vesicles following 36 h expression. D) Counts of vesicle co-expressing ABCA3WT•DsRed and EGFP•hSP-C isoforms. Data were obtained from three separate experiments, each with counts of 50–100 vesicles per cell in at least 100 transfected cells. *p < 0.001, **p < 0.01, ***p < 0.05.

Mistargeted hSP-CI73T is also aberrantly processed

In alveolar type II epithelial cells, SP-C is synthesized as a 21-kDa bitopic integral membrane precursor that undergoes four proteolytic cleavages as it is trafficked through the biosynthetic pathway to yield a 3.7-kDa mature peptide (2). However, in A549 and other cell lines, only the initial of these four cleavages occurs at or near residue 145 in early post-Golgi compartments (Figure 1) (24,28,29). Moreover, epithelial cell lines such as A549 or HEK293 secrete neither the mature nor the processed forms of SP-C. However, these cell lines have been proven advantageous for their own benefits over primary alveolar type II cells since the primary cells are phenotypically unstable in culture.

Western blot analyses of EGFP•hSP-C isoforms in transfected A549 cell lysates using antisera directed against the N-terminally placed EGFP tag demonstrated that proSP-CI73T undergoes abnormal post-translational processing. As shown in Figure 4A, a major product of 27 kDa was detected in cells transfected with EGFP alone. When wild-type EGFP•hSP-C was introduced, lysates contained two anti-EGFP-positive bands that included the primary translation product of the fusion protein (48 kDa) (Figure 4A, arrowhead) and a smaller processed intermediate consistent with previous reports of partial C-terminal processing of normally routed proSP-C by A549 cells (10,24). Contrary to wild-type proSP-C, expression of EGFP•hSP-CI73T resulted in multiple mutant proSP-CI73Tproducts composed of the primary translation and processed intermediate forms (similar to those observed in the wild type isoform) as well as two additional forms (Figure 4A, arrows), one migrating at higher Mr (∼50 kDa) and the other, with Mr between the primary translation and processed intermediate products. The lower product migrated at a higher molecular weight than the normally processed product suggesting that the cleavage site is in the C-terminus and more distal than the normal cleavage site. Treatment with PNGase F, O-glycosidase or protein phosphatases did not alter the migration of any of the proSP-CI73Tproducts (data not shown).

Figure 4.

EGFP•hSP-CI73T is trafficked to the plasma membrane. A) Representative immunoblotting (using EGFP antibody) of whole-cell lysates of nontransfected (Blank), vector transfected (EGFP-C1), or transfected with EGFP•hSP-CWT or EGFP•hSP-CI73T. The primary translation products (arrowhead) and processed products (bottom bands) of the hSP-C isoforms are shown. Two aberrantly processed products in EGFP•hSP-CI73T cell lysates (arrows) are apparent. β-actin bands were used to assess equal loading. B) Representative epifluorescent images of A549 cells following an overnight expression of either EGFP•hSP-CWT (top panel) or EGFP•hSP-CI73T (bottom panel) and cell surface labeling of plasma membrane proteins with biotin–streptavidin complex at 4°C according to protocols described in Materials and Methods. The complex was visualized using Texas Red-conjugated streptavidine antibody. Bar, 5 µm. C and D) A549 cells grown in T75 flask were transiently transfected and cell surface labeled with biotin–streptavidin as in (B). C) Representative immunoblotting (using anti-EGFP) of plasma membrane-bound proteins following immunoprecipitation using anti-streptavidin. Unlike EGFP•hSP-CWT, the aberrantly processed products (arrows) in addition to the primary translational product (arrowhead) of the EGFP•hSP-CI73T protein were preferentially localized at the plasma membrane. Biotinylated proteins were normalized by loading equal concentration of each protein sample (Bradford method (30)). D) Band intensities of plasma membrane expressed EGFP-tagged total SP-C products were quantified and results from at least three separate experiments are shown. *p < 0.001.

hSP-CI73T is preferentially targeted to plasma membrane

In addition to EEA1 vesicles, a significant fraction of EGFP•hSP-CI73T fluorescence appeared to be localized to plasma membrane (Figure 2, arrows). To confirm the appearance of EGFP•hSP-CI73T at the plasma membrane, surface biotinylation of intact transfected cells was performed at 4°C. Compared to EGFP•hSP-CWT, immunofluorescence imaging of EGFP•hSP-CI73T transfected A549 cells showed co-localization of the mutant isoform with biotin–Texas-Red–streptavidin conjugates on the plasma membrane (Figure 4B, arrows). Analysis of biotinylated cell surface protein by streptavidin pull down followed by Western blot, provided biochemical confirmation that significant amounts of the EGFP•hSP-CI73T primary translation product reached the plasma membrane (Figure 4C (arrowhead),D). Interestingly, the two aberrantly processed forms of hSP-CI73T also appeared on the cell surface (Figure 4C, arrows).

hSP-CI73T colocalizes with transferrin both at the plasma membrane and in early endosomal compartments

To link the observation of cell surface expression and localization of EGFP•hSP-CI73T to early endosomes, time-dependent analysis of mutant propeptide trafficking following plasma membrane localization was performed using Alexa Fluor-conjugated transferrin as a plasma membrane and endocytosis marker. In the absence of endocytosis (4°C), EGFP•hSP-CI73T co-localized with Alexa Fluor-transferrin at the cell surface (Figure 5B, top row). Increasing the temperature to 37°C for 5 min consistently resulted in internalization of transferrin with subsequent co-localization in EGFP•hSP-CI73T containing vesicles primarily in proximity to the plasma membrane (Figure 5B, middle row). Further incubation at 37°C for 1 h resulted in further co-localization in vesicles distributed throughout the cell (Figure 5B, bottom row). In contrast, EGFP•hSP-CWT predominantly remained within cytosolic vesicles and rarely colocalized with transferrin at any of the time points examined (Figure 5A).

Figure 5.

Time-dependent colocalization of EGFP•hSP-CI73T and transferrin at the plasma membrane and early endosomal vesicles. Representative epifluorescent images of live A549 cells following 24 h expression of wild type and mutant EGFP-tagged hSP-CWT (A) and hSP-CI73T (B) and incubation with Alexa Flour 594-labeled transferrin at 4°C for 1 h and fixed (top rows) or transferred and incubated at 37°C for 5 min (middle rows) and 1 h (bottom rows). Plasma membrane co-localizations of EGFP•hSP-CI73T with transferrin (arrows) followed by co-localization of the two proteins initially confined in vesicles proximal to the plasma membrane (arrowheads) after a 5-min incubation at 37°C (B, merged middle image), and their subsequent distribution throughout the cell (B, merged bottom image) after 1-h incubation at 37°C are shown. Bar, 5 µm.

hSP-CI73T is initially routed directly to plasma membrane

While the current data demonstrate the presence of SP-CI73T at the cell surface, there are several possible routes that could be utilized for initial delivery of the protein to plasma membrane (31). To investigate whether the mutant protein is trafficked via Golgi to endosome to plasma membrane pathway, or is directly delivered to the cell surface, time-dependent colocalization studies were performed utilizing the mannose-6-phosphate receptor (M6PR) as an endosome marker, or wheat germ agglutinin (WGA) as a plasma membrane marker in A549 cells expressing EGFP-tagged hSP-CI73T. At earlier time points (6 and 12 h), faint EGFP•hSP-CI73T expression was observed that only sparsely colocalized with M6PR (Figure 6A,B) and did not reach a maximum until 24 h post-transfection (Figure 6B). In contrast, significant early colocalization of the mutant isoform with WGA was observed starting at 6 h post-transfection and peaking as early as 16 h post-transfection (Figure 6A,B). These results suggest that the initial route taken by hSP-CI73T to the plasma membrane is through a direct, anterograde, nonendosomal pathway.

Figure 6.

EGFP•hSP-CI73T is initially trafficked to the plasma membrane via a nonendosomal pathway and subsequently internalized. A) Representative confocal fluorescent images of permeabilized and labeled with anti-M6PR (left panel), or nonpermeabilized and labeled with WGA (right panel) A549 cells expressing EGFP•hSP-CI73T 6, 12, 16, 20, and 24 h following transfection. Bar, 5 µm. Due to low level of fluorescent expression, visualization at 6 and 12 h post-transfection was accomplished by significantly increasing confocal device settings (“iris size”, “gain”, and “offset” was increased up to 4, 4, and 3 times above normal, respectively). B) Cell count of expressed EGFP•hSP-CI73T colocalized with markers. For M6PR, quantitation was performed by counting cells that showed colocalization of M6PR in at least three EGFP•hSP-CI73T-containing vesicles. At least 50 cells per marker from two separate experiments were counted. *p < 0.05, **p < 0.01. C) Representative confocal fluorescent images of live A549 cells following 24 h expression of wild type and mutant EGFP-tagged hSP-CWT (top row) and hSP-CI73T (bottom two rows) and incubation with epitope specific C-terminal proSP-C antibody at 4°C for 1 h and fixed (middle row) or transferred and incubated at 37°C for 30 min (bottom row). Solid boxes from the merged images are magnified to illustrate localization or nonlocalization (right columns). Membrane localizations of EGFP•hSP-CI73T (arrowheads) followed by internalization in punctate vesicles (arrows) after 30-min incubation at 37°C are shown. Images are representative of at least three separate experiments. Bars, 5 µm.

hSP-CI73T is internalized from plasma membrane to cytosolic vesicles

Although intracellular co-localization of transferrin with SP-CI73T was apparent following transferrin internalization, there was no direct evidence demonstrating whether the SP-CI73T found in transferrin-containing organelles was actually from the plasma membrane. To substantiate SP-CI73T endocytosis, an epitope-specific antisera raised against the C-terminal region of the propeptide (anti-CTERM) (32) was used. As anticipated, in the absence of endocytosis (4°C), EGFP•hSP-CI73T was localized at the cell surface (Figure 6C, middle row, arrowheads). Increasing the temperature to 37°C for 30 min consistently resulted in the internalization of the antibody-labeled EGFP•hSP-CI73T to punctate donut-shaped vesicles (Figure 6C, bottom row, arrows). In contrast, EGFP•hSP-CWT predominantly remained within cytosolic vesicles did not reach the plasma membrane and was not labeled by α-CTERM (Figure 6C, top row).

Expression patterns of hSP-CI73T in primary human alveolar type II cells are similar to epithelial cell lines

Primary alveolar type II cells were isolated (from human fetal lung in limiting quantities) and transfected in order to verify SP-CI73T expression patterns observed in A549 and HEK293 (Figure 8) cell lines. Confocal immunofluorescence imaging of transiently transfected alveolar type II cells showed distinct distribution of the wild-type (Figure 7, left panel) versus the mutant (Figure 7, right panel) EGFP•SP-C fusion proteins. As expected, wild-type SP-C vesicles were double positive for CD63 and DC-Lamp (another marker antigen for human alveolar type II cell lamellar bodies and lysosomes) and negative for EEA1. In contrast, the EGFP•hSP-CI73T mutant isoform was preferentially expressed in compartments that were positive for EEA1 but also positive for CD63 and DC-Lamp likely secondary to the homomeric/heterotypic association with endogenous SP-C in these cells (see Results, first paragraph). In addition, EGFP•hSP-CI73T but not EGFP•hSP-CWT was found on the plasma membrane (arrowhead) co-localizing with the plasma membrane marker, epithelial cell adhesion molecule (EpCAM) (arrow).

Figure 8.

Lipid uptake and degradation are impaired by hSP-CI73T expression. A) Representative confocal fluorescent images of nonpermeabilized HEK293 cells stably expressing EGFP-tagged wild type (top raw) or I73T (bottom raw) isoforms of SP-C labeled with the plasma membrane marker, WGA. Plasma membrane colocalization of the SP-CI73Twith WGA is apparent with merged (yellow) images (bottom right). Bar, 5 µm. B) Representative anti-GFP immunoblotting bands (top panel) and histogram of quantified bands (bottom panel) of HEK 293 whole-cell lysate samples stably expressing either EGFP•hSP-CWT (middle lane) or EGFP•hSP-CI73T (right lane) are displayed to demonstrate equal expression of SP-C isotypes. Nontransfected HEK cells were used as control (left lanes). β-Actin was used to normalize for equal loading. C and D) Histograms of radioactivity measured as disintegration per minute (DPM) of DPPC uptake per 0.5 million cells (C) and % DPPC degradation (D) in HEK293 cells treated with H3-DPPC-containing liposomes stably expressing EGFP•hSP-CWT (middle lane) or EGFP•hSP-CI73T (right lane). Data were obtained from at least three separate experiments for (A) and (B), and at least seven separate experiments for (C) and (D) each consisting of at least two culture dishes per construct. *p < 0.01; **p < 0.05.

Figure 7.

Sorting of hSP-CI73T in primary human alveolar type II cells is similar to epithelial cell lines. Representative confocal images, from at least three separate experiments, of primary alveolar type II cells isolated from human fetal lung tissues transiently expressing EGFP-tagged wild type (left panel) or mutant I73T (right panel) SP-C isoforms immunostained with Texas Red-tagged subcellular markers. Both isoforms were found in compartments containing the lysosome/lamellar body markers DC-lamp (top row) and CD63 (second row) (depicted in yellow merged images). Bottom two rows, co-localization of mutant I73T isoform with early endosome and plasma membrane markers EEA1 (third row) and Ep-CAM (bottom row), but not with wild-type SP-C. Solid boxes from the merged images are magnified to illustrate colocalization or noncolocalization (right columns). Distinct plasma membrane localization of the mutant isoform is apparent (second row, arrowhead and bottom row, arrow). Bar, 5 µm.

Expression of hSP-CI73T impairs cellular uptake and degradation of lipids

In addition to parenchymal remodeling, patients with hSP-CI73T mutation exhibit alveolar phospholipidosis (11,33) suggesting a defect in surfactant lipid metabolism. On the basis of the mistargeting of hSP-CI73T, we hypothesized that cells expressing the mutant isoform would exhibit perturbation of cellular processes that are dependent upon endosomal–lysosomal progression which would include lipid metabolism.

To examine the functional consequences from expression of the mutant hSP-C protein, lipid uptake and degradation were generated from clones of HEK293 cells stably expressing either EGFP tagged wild type or mutant hSP-C isoforms (Figure 8A). As for A549 cells, immunofluorescence analysis of the stable cell lines showed a similarly distinct expression pattern where plasma membrane localization of the mutant isoform with WGA was apparent (Figure 8A, bottom row). By Western blot, the HEK293 clones demonstrated equal expression of EGFP•hSP-CWTor EGFP•hSP-CI73T (Figure 8B).

Cell monolayers were incubated at 37° with H3-choline-labeled DPPC-containing liposomes, and uptake was measured 2 h following treatment. As demonstrated in Figure 8C, DPPC uptake was significantly reduced in cells expressing EGFP•hSP-CI73T (29% reduction) compared to those expressing EGFP•hSP-CWTor nontransfected cells. Moreover, degradation of DPPC that had been internalized by the cells during the 2-h incubation period was differentially altered (Figure 8D). In cells expressing EGFP•hSP-CWT, degradation averaged 5.5% of internalized DPPC. In contrast, degradation of DPPC in EGFP•hSP-CI73T expressing cells was reduced by more than a third (36% reduction) or by 25% compared to cells expressing EGFP•hSP-CWT or nontransfected cells, respectively. Taken together, these results suggest that expression of hSP-CI73T functionally impairs lipid uptake and degradation by epithelial cells.


The cellular mechanisms underlying the pathogenesis of ILD associated with SFTPC mutations are partly understood. Most of the published reports have focused on a select group of mutations located in the proprotein domain flanking the SP-C C-terminus known as the BRICHOS domain (34) where the unfolded protein response (UPR), ER stress, mitochondrial dysfunction and apoptosis have been implicated (7,9,10). In contrast to these, expression of two different non-BRICHOS mutations (I73T, E66K) of the SFTPC gene produces a distinct clinical as well as cellular phenotype. In vivo, non-BRICHOS mutants produce both interstitial pneumonia and pulmonary lipoproteinosis accompanied by abnormal localization of the SP-C propeptide in EEA-1 vesicles in alveolar type II cells (11,35). Since these reports, the hSP-CI73T mutation is now recognized as the most common SP-C mutation to date (13,14,16,17,36). The present report extends the prior findings by demonstrating that trafficking of the hSP-CI73T mutant is altered by its misrouting to the plasma membrane, subsequent internalization to early endosomes and delayed appearance within lysosomal-like compartments. This misrouting is accompanied by aberrant post-translational processing of the mutant isoform, and these abnormalities collectively result in functional impairment of lipid uptake and degradation. Thus, the hSP-CI73T non-BRICHOS mutation represents a new class of spatially and functionally distinct mutations of proSP-C that is capable of producing chronic lung disease, but is phenotypically different from those of misfolded BRICHOS mutants.

In general, the diversion of a protein from reaching its targeted destination due to missense, nonsense, deletion or frame shift mutations is a common phenomenon that occurs in diseases associated with secretory or transmembrane protein mutations. In many cases, protein misfolding is believed to contribute to the pathophysiology. Best known among these are neurodegenerative diseases including Huntington's disease, Alzheimer's disease, amyotrophic lateral sclerosis and Parkinson's disease. For each, a gain-of-function results from mutations in Huntingtin, beta amyloid, SOD1 and α-synuclein, respectively, and produces ER retention of the misfolded protein leading to the activation of the UPR, ER stress and protein aggregation (37). These cellular responses in turn lead to the subcellular disruption of essential proteins such as chaperones, interfere with degradation of other proteasome substrates and compromise the folding and stability of other proteins that are conformationally challenged to cause the global disruption of cellular homeostasis leading to cell death. We have shown some of these cellular abnormalities also occur in vitro in ILD-associated BRICHOS domain hSP-C mutants, including deletion of the exon 4 transcript of the hSP-C gene (Δexon 4) and a missense mutation that substitutes leucine for glutamine at residue 188 (L188Q) (6,10,38). The absence of mature hSP-C in the bronchoalveolar lavages of patients with the heterozygous expression of these mutations suggests a dominant negative effect by the mutant isoform.

Utilizing multiple experimental techniques, data in the present study demonstrate that the hSP-CI73T mutant propeptide is misrouted to the plasma membrane. In contrast to the BRICHOS mutants, hSP-CI73T fails to induce either a dominant negative effect as some mature SP-C can be detected in patients heterozygous for hSP-CI73T or a classic toxic gain of function as significant amounts of ER stress are lacking (9,35). In this way, hSP-CI73T-induced ILD may be more similar to cytotoxic degenerative diseases such as primary hyperoxaluria type 1 (PH1) disease (39), malonyl-coA decarboxylase (MLYCD) deficiency (40) and hereditary hypophosphatemic rickets with hypercalciuria (HHRH) (41). These diseases all result from misrouting of missense mutants to different subcellular compartments or to the plasma membrane, and in most cases lead to cellular dysfunction. Complementary to this concept, our results suggest expression of the aberrantly targeted mutant hSP-CI73T produces a secondary loss-of-function via disruption of lysosomal degradation processes. Moreover, the functional impairment need not be limited to lipid uptake and degradation, but may also extend to other proteins that share these routing compartments. Aberrant processing intermediates of SP-B have been detected in intra-alveolar surfactant fractions in patients with both hSP-CI73T and hSP-CE66K mutations (11,19,35).

In the present study, we have also detected aberrant EGFP•hSP-CI73T intermediates which likely result from mistargeting-dependent exposure of the mutant isoform to alternative processing events from different or additional protease(s). As shown in Figures 4 and 6, there are two additional bands that are absent in the expression profile of wild-type EGFP•hSP-C isoform. These have also been recently identified both in the culture media of cells stably expressing hSP-CI73T and in the lungs of patients with this mutation (19,28). The first band, migrating faster than the primary translation product, could result from intracellular enzymatic cleavage caused by the propeptide's improper routing to EEA1 vesicles. Alternatively, since the same product was biotinylated (Figure 4C), it is possible that trimming of the C-terminal propeptide at the cell surface by opportunistic extracellular proteases may occur. As shown in Figure 9, since proSP-C is a type II integral transmembrane protein (32,42), direct delivery to the cell surface would result in a membrane spanning orientation whereby the C-terminal region is exposed to the extracellular milieu making alternative potential protease cleavages a spatial possibility. Biochemical studies of patients with hSP-CI73T as well as the E66K mutation support the findings of abnormal processing. Aberrantly processed 11- and 13-kDa C-terminal segments of hSP-C proprotein were identified by immunoblotting of bronchoalveolar lavage fluid of a patient carrying the SP-CI73T mutation (11,35). Mass spectrometry has also shown intra-alveolar accumulation of aberrant C-terminal processing products (19).

Figure 9.

Model for aberrant hSP-CI73T trafficking. Schematic representation depicting normal and abnormal sorting of wild-type and mutant isoforms of SP-C. The hSP-CWT gene/cDNA is transcribed into a 0.9-kb mRNA product which is translated to a 197 amino acid (21 kDa) proprotein. Proteolytic cleavages occur during transit through the secretory pathway that include small vesicles, multivesicular bodies and lysosome-like vesicles (LLV). In contrast, hSP-CI73T is misdirected to the plasma membrane via a nonendosomal and likely through the constitutive pathway where it is subsequently internalized and sorted to early endosomes and LLV. N and C depict the N- and C-termini of the proprotein. Hatched arrows imply pathways that may require multiple steps to reach target organelles. (Inset) Topological representation of the bitopic proSP-C with-type II orientation. Mature SP-C (blue), cleavable segments of proSP-C (green), Juxtamembrane palmitoylation sites (arrow), and a C-terminal disulfide-bonding site (arrowhead) are shown.

In addition to abnormal protein processing, other consequences of hSP-CI73T mistargeting include an accompanied ultra structural dysfunction. Alveolar type II cells are actively involved in the clearance of pulmonary surfactant from the intra-alveolar space as well as recycling [via the multivesicular bodies (MVBs)] of internalized surfactant lipids and proteins back to the lamellar body (43,44). Interestingly, the lamellar body itself exhibits characteristics of a lysosomal-like secretory organelle (45). Consequently, the misrouting of proSP-CI73T to the cell surface and its subsequent internalization not only affects the lysosomal degradation pathway, but may also disrupt the MVB/lamellar body recycling system including docking and fusion processes of exocytosis. Thus, the observation of abnormal dense core organelles and poorly developed lamellar bodies in alveolar type II cells of patients with hSP-CI73T(11,17) is likely the product of both disrupted lysosomes and a functionally compromised endosome/MVB/lamellar body recycling system. Moreover, since patients with SP-CI73Thave alveolar lipoproteinosis, the increased steady-state level of surfactant reflects impaired endocytosis (supported by the present study) and also implies enhanced secretion.

To address this functionally, we modeled the effect of hSP-CI73T expression on clearance of lipid components using stably transfected cell lines. The reduced phospholipid uptake and degradation demonstrated in Figure 8 conforms to disorders associated with defects in clearance and recycling of surfactants from lung. Under normal conditions, both type II cells and alveolar macrophages contribute to surfactant clearance from the distal airways. Previously, in both humans and transgenic mice, defects in granulocyte-macrophage colony-stimulating factor (GM-CSF) signaling either by ligand- or receptor-inhibition (by neutralizing antibodies or mutations) have been associated with disruption of surfactant homeostasis where the bronchoalveolar lavages of patients lacking GM-CSF activity have revealed pulmonary alveolar lipoproteinosis (46–48). Likewise, targeted deletion in mice of the acid-independent phospholipase A2(aiPLA2), a lysosomal enzyme that degrades endocytosed surfactant DPPC in alveolar type II cells, results in the time-dependent accumulation of both total surfactant phospholipid [mainly disaturated phosphatidylcholine (PC) in bronchoalveolar lavage, lung homogenate and lung lamellar bodies] (49). Therefore, the intra-alveolar accumulation of lipids and protein in patients with hSP-CI73T mutation could directly be attributed to a malfunction in type II cell-dependent surfactant clearance and recycling and suggests both cell types are important for surfactant homeostasis.

In summary, we report phenotypic and biochemical characterization of a non-BRICHOS mutation of the SFTPC gene associated with interstitial pneumonia and alveolar lipoproteinosis. We have determined aberrant protein trafficking and subcellular localization (Figure 9) of this propeptide leading to abnormal protein processing. The nature of the mutation appears to dictate an alternative biosynthetic route taken by the mutant propeptide, which is distinctly different in both spatial and temporal phases from those observed within BRICHOS folding mutants. The acquired impairment of lipid uptake and degradation by cells expressing the mutant propeptide support data from earlier reports of patients with abnormal intra-alveolar contents of lipids and proteins, including increased surfactant protein A content and aberrantly processed surfactant proteins B and C. In addition, this functional impairment points to a novel concept where a protein that is not functionally associated with lysosomes can compromise the lysosomal-dependent degradation process. Understanding molecular mechanisms underlying the etiology and evolution of these abnormalities will provide a pertinent perspective to disease progression, and may offer insights into the development of new therapeutic strategies tailored to the specific subcellular deficits and pathways induced by such mistargeted proteins.

Materials and Methods


The pEGFP-C1 and DsRed-N1 monomer plasmids were purchased from Clontech, Inc. Tissue culture medium was produced by the Cell Center Facility, University of Pennsylvania. Except where noted, all other reagents were electrophoresis, tissue culture or analytical grade and were purchased from Sigma Chemical, Inc. or BioRad, Inc.

Generation of plasmid constructs

Constructs used for this study are illustrated in Figure 1. Human wild-type proSP-C (amino acids 1–197) inserted into the pEGFP-C1 vector has been previously described, as has EGFP•hSP-CI73T(24,38). A full-length ABCA3WT•DsRed was generated by substitution of the EGFP fluorescence encoding region of the ABCA3WT•EGFP construct (26) with nondimerizing DsRed encoding region using restriction enzymes XhoI and EcoRI.

Cell lines and transfection

Human A549 and HEK293 epithelial cell lines were originally obtained through the American Type Culture Collection (ATTC). A549 cells grown to 85–90% confluency on glass coverslips in 35-mm plastic dishes were transiently transfected with various fusion wild-type or mutant hSP-C constructs (4 µg/dish) or co-transfected with post-Golgi distal vesicle marker construct, hABCA3WT•DsRed (2 µg/dish per construct), using Lipofectamine 2000 (Invitrogen). EGFP•hSP-CWT and EGFP•hSP-CI73T stable cell lines in HEK293 cells were generated following a 48-h transient expression using CaPO4 transfection method (32), and with subsequent steps taken for clonal selection with 1 mg/mL G418 as described previously (50–52). The cells were maintained in Dulbecco's Modified Eagles medium (DMEM) containing 500 µg/mL G418, 10% FBS and 1% penicillin/streptomycin.

Human fetal lung alveolar type II cell culture and transfection

Isolation and culture of human fetal alveolar type II cells were performed as previously described (53). Briefly, enriched populations of epithelial cells were isolated from second-trimester (13–20 weeks' gestation) human fetal lung tissue obtained under a protocol approved by the Children's Hospital of Philadelphia Institutional Review Board. After overnight culture as explants without hormones, the tissue was digested with trypsin, collagenase and DNAase and removed fibroblasts by differential adherence. Nonadherent cells were first transfected with various wild-type or mutant SP-C fusion constructs using electroporation with the Amaxa transfection system prior to plating on glass cover slips in 30-mm plastic culture dishes in Weymouth's medium containing 10% fetal calf serum. After overnight culture (day 1), attached cells were cultured for an additional 3–9 days in 1 mL of serum-free Weymouth's medium containing 10 nM dexamethasone, 0.1 mm 8-bromo-cAMP and 0.1 mm isobutylmethylxanthine together (referred to as DCI), previously shown to maximally induce surfactant components consistent with a morphological alveolar type II cell phenotype.

Sucrose gradient fractionation

Subcellular fractionation using discontinuous sucrose gradient was performed using the method of Stockinger et al. with some modifications (54). A549 cells in T75 flasks transiently expressing EGFP-tagged hSP-C isoforms for 24 h were harvested by scraping and were washed twice with phosphate-buffered saline (PBS) (137 mm NaCl, 10 mm Na2HPO4, 2.7 mm KCl, 1.8 mm KH2PO4, pH 7.4). Cells were subsequently resuspended with 750 µL of 0.2 m sucrose containing protease inhibitors [1 mm phenylmethylsulfonyl fluoride (PMSF), leupeptin (1 µg/mL), pepstatin A (1 µg/mL) and aprotinin (1.5 µg/mL)] and disrupted by 10 strokes using a glass tissue homogenizer and Teflon probe. Nuclei were removed by 5-min centrifugation at 1000 ×g and the supernatant was adjusted to 750 µL with 0.2 m sucrose. The supernatant was loaded on top of 5.25 mL containing 0.4, 0.6, 0.8, 0.9 and 1.0 m discontinuous sucrose gradient (750 µL each) and spun for 1 h in a Sorvall ultracentrifuge (swinging bucket rotor, TH-641) at 105 000 ×g. Individual fractions (750 µL) were collected from the bottom up using Pasteur pipettes and sample proteins were concentrated using trichloroacetic acid (TCA) precipitation method comprised of incubating samples on ice in equal volume of 20% trichloroacetic acid (TCA) for 30 min; centrifugation at 13 200 ×g for 15 min; washing precipitates twice with 500 µL ice cold acetone; and resuspending samples in SDS/PAGE loading buffer.

Cell surface biotinylation

Cell surface protein biotinylation was performed using the Sulfo-NHS–SS-Biotin reagent with Streptavidin Agarose Resins according to the manufacturer's protocol (Thermo Scientific) with some modifications. A549 cells transiently expressing isoforms EGFP•hSP-C in T75 flask grown to 85–90% confluency were washed twice with 8 mL ice-cold PBS, pH 7.4. A 10 mL–10 mm of Sulfo-NHS–SS-Biotin was added, respectively, to the flask and gently agitated for 30 min at 4°C. The biotinylation reaction was stopped by adding 500 µL of quenching solution (Thermo Scientific). Cells were harvested by scraping and washed three times by centrifugation in Tris-buffered saline (TBS) (20 mm Tris, 500 mm NaCl, pH 7.5) at 500 ×g for 3 min. The resulting pellet was resuspended in 400 µL HEN buffer (25 mm HEPES, 0.1 mm EDTA, 10 µm Neocuporine) containing protease inhibitors and 1% SDS (added fresh). Cells were sonicated using a probe sonicator (60 Sonic Dismembrator, Fisher Scientific) utilizing three burst of 10 seconds with a 15-second interval and at 40% maximum power output. Streptavidin binding was accomplished by adding 200 µL neutralization buffer (20 µm HEPES, 1 mm EDTA, 10 µm NaCl, 0.5% Triton X-100) and 500 µl of streptavidin–agarose beads (Thermo Scientific) to the sonicated solution and incubating at room temperature with end to end rotation overnight. The bound biotin–streptavidin–agarose complex was washed five times using 500 µL neutralization buffer containing 600 mm NaCl. The agarose beads were released by adding 20 µL 2× lithium dodecyl sulfate (LDS) sample buffer (Novex), 2 µL 10× sample reducing agent (Novex), and by heating at 95°C for 10 min.


Cell pellets collected from 35-mm culture dish by scraping and centrifugation at 300 ×g were solubilized with 40 µL of lysis buffer containing protease inhibitors. Following centrifugation at 6,000 ×g for 30 seconds to remove nuclei, proteins were separated by electrophoresis on a 12% polyacrylamide gel and transferred to nitrocellulose membrane.


Immunoblotting of transferred samples was performed using successive incubations with primary anti-EGFP (Clontech), 1:5000; anti-Lamp 1 (Abcam Inc.,), 1:1000; anti-EEA1 (BD Biosciences), 1:3000; anti-β-actin (Sigma), 1:2000; and horseradish peroxidase-conjugated secondary antibodies as previously described (24). The resulting fluorescence images visualized by ECL were captured either by exposure to film or by direct acquisition using the Kodak 440 Imaging System.


Colocalization studies were performed by immunostaining cells plated on coverslips and fixed by coverslip immersion in 4% paraformaldehyde as previously described (9). Following permeabilization, cells were immunolabeled with primary antibodies for 1 h at room temperature at the following dilutions: anti-CD63 (Immunotech), 1:500; anti-EEA1 (BD Biosciences), 1:3000; anti-streptavidin 1:1000, (Abcam Inc.), DC-Lamp (Beckman Coulter) 1:100; Ep-CAM (Santa Cruz Biotech) 1:200; anti-CTERM (Beers Laboratory), 1:250; anti-M6PR (Abcam Inc.), 1:150. Texas red-conjugated secondary goat anti-mouse IgG monoclonal or secondary goat anti-rabbit IgG polyclonal antibodies (Jackson ImmunoResearch Laboratories) were used for visualization at a dilution of 1:2000.

Plasma membrane labeling with WGA

Cells plated on coverslips were fixed by coverslip immersion in 4% paraformaldehyde and without permeabilization. Cells were washed 3 × 5 min with PBS, and 3.3 µL/mL of 1 mg/mL Alexa Fluor 594 WGA (Invitrogen Molecular Probes) stock solution was added to cells and incubated with gentle agitation for 30 min. Cells were subsequently washed 3 × 5 min with PBS and cover slips mounted on slides for fluorescence visualization.

Fluorescence microscopy

EGFP images of permeabilized/nonpermeabilized cells fixed in 4% paraformaldehyde were viewed on an Olympus I-70 inverted fluorescence microscope using a High Q FITC filter packages (Excitation 480 nm; Emission 535/550 nm) (Chroma Technology). Similarly, DsRed images and Texas Red-tagged immunostained images were visualized with High Q TR filter (excitation, 560/555 nm; emission, 645/675 nm) as described (55,56). Image acquisition, processing and overlay analysis were performed using the Metamorph 7.5 software (Molecular Devices, Inc.) by utilizing identical imaging criteria within experimental groups including equal exposure time, image scaling and image gamma. For confocal microscopy, cells were examined using the TE300 Nikon coupled Radiance 2000 imaging system, Carl Zeiss, by using identical imaging criteria within experimental groups including equal exposure time, image scaling and image gamma. All Images were subsequently transformed from 24 bit program files to 8 bit TIFF files and transferred to a Power Point program for assembly and labeling. When necessary, enhancement of images was performed during the Power Point assembly by adjusting the “brightness” and “contrast” of each image within the same figure equally.

Dynamic trafficking of EGFP•hSP-C isoforms

Following 24 h expression of EGFP•hSP-C isoforms in A549 cells grown on coverslips in 35 mm culture dishes, cell medium was replaced by fresh, ice-cold culture medium containing 20 µg/mL Alexa Flour 594-labeled transferrin (Invitrogen) or epitope-specific C-terminal polyclonal antibody (anti-CTERM) (32). Cells were incubated with gentle agitation for 60 min at 4°C. Cells were washed with ice-cold culture medium to remove excess transferrin or antibody from the cell surface and either washed with ice cold PBS and fixed directly, or incubated at 37°C for 5, 30 or 60 min, washed in cold PBS, and then fixed. For fluorescence visualization, the cover slips from the transferrin experiment were directly mounted on slides and the cover slips from anti-CTERM binding were subsequently incubated with Texas Red-conjugated secondary antibody prior to slide mount.

Preparation of liposomes

Unilamellar liposomes were prepared by constituting 0.5 m DPPC, 0.25 m egg PC, 0.15 m cholesterol and 0.1 phosphatidylglycerol to reflect the approximate composition of lung surfactant (49,57). Radiolabeled DPPC containing [3H-methyl] choline was added to a final concentration of 0.052 µci/µL. Unilamellar liposomes with mean diameter of 108 nm were prepared by evaporating the mixture of lipids under N2 to dryness, resuspending the evaporated film in PBS with vigorous mixing, freezing and thawing three times by alternating exposure to liquid N2 and a 50°C water bath, followed by extrusion of the sample at 50°C for 10 cycles through a 100-µm pore size filter.

Uptake and degradation of DPPC

Uptake of 3H-DPPC in liposomes by HEK293 cells or by HEK293 cells stably expressing either wild-type or mutant isoforms of hSP-C was measured based on methods described previously (49,58,59) with some modifications. Growth medium was removed from HEK293 cells grown to confluency in 35 mm dishes and was replaced with DMEM (1 mL/dish) containing 20 nM radiolabeled unilamellar liposomes. Following incubation at 37°C for 2 h in a 5% CO2 incubator, cells were sequentially washed twice with DMEM, once with PBS and removed from dishes by trypsinization. Three dishes per sample were pooled for analyses. Cells were pelleted by centrifugation and were resuspended in 0.5 mL PBS followed by cell counts using a hemocytometer. Samples were normalized for equal number of cells per sample by dilution with PBS. Each sample was then equally divided and one half was directly analyzed for radioactivity measured as disintegration per minute (dpm) to determine total DPPC uptake. Uptake of DPPC was calculated from total DPM measured per 0.5 million cells. The remaining half of the sample was fractionated using the Bligh and Dyer method (60) and the amount of DPPC degradation was calculated as the percentage of radioactivity in the aqueous fraction relative to the total radioactivity present in the lipid plus aqueous fractions.

Data analysis

Experimental data were analyzed by one-way analysis of variance (anova) with the Tukey-Kramer post hoc test using graphpad instat software, version 3.0 for Windows (GraphPad Software). All values are means ± SE. Significance was accepted at p < 0.05.


We thank Dr. Aron B. Fisher, Chandra Dodia and Jain-Qin Tao of the University of Pennsylvania, Institute for Environmental Medicine, for their suggestions and technical assistance with lipid assays. This work was supported by National Heart, Lung and Blood Institute Grants HL09073 (S. M.), HL059959 (S. G.), HL19737 (M. F. B.) and HL087177 (M. F. B.).