An 11-nt sequence polymorphism at the 3′UTR of human SFTPA1 and SFTPA2 gene variants differentially affect gene expression levels and miRNA regulation in cell culture

Abstract

Surfactant protein A (SP-A) plays a vital role in maintaining normal lung function and in host defense. Two genes encode SP-A in humans (SFTPA1, SFTPA2), and several gene variants have been identified for these. We have previously shown that sequence elements of SFTPA1 and SFTPA2 3′ untranslated regions (UTRs) differentially affect translation efficiency in vitro. Polymorphisms at the 3′UTRs of mRNA variants may account for differential binding of miRNAs, a class of small noncoding RNAs that regulate gene expression. In this work, we generated 3′UTR reporter constructs of the SFTPA1 and SFTPA2 variants most frequently found in the population, as well as mutants of a previously described 11-nt indel element (refSNP rs368700152). Reporter constructs were transfected in NCI-H441 cells in the presence or absence of miRNA mimics, and reporter gene expression was analyzed. We found that human miRNA mir-767 negatively affected expression of constructs containing SFTPA1 and SFTPA2 variants, whereas mir-4507 affected only constructs with 3′UTRs of SFTPA1 variants 6A, 6A³, and 6A⁴ (not containing the 11-nt element). Three miRNAs (mir-183, mir-449b, and mir-612) inhibited expression of recombinants of SFTPA2 variants and the SFTPA1 variant 6A², all containing the 11-nt element. Similar results were obtained for SP-A expression when these miRNAs were transfected in Chinese hamster ovary cells expressing SFTPA1 or SFTPA2 variants or in NCI-H441 cells (genotype 1A⁵/1A⁵-6A⁴/6A⁴). Moreover, transfection with a specific antagomir (antagomir-183) reversed the effects of mir-183 on SP-A mRNA levels. Our results indicate that sequence variability at the 3′UTR of SP-A variants differentially affects miRNA regulation of gene expression.

human surfactant protein a (SP-A) is encoded by two homologous genes, SFTPA1 and SFTPA2. These are located in chromosome 10 in opposite transcriptional orientation (5, 25). The proteins encoded by these two genes (SP-A1 and SP-A2) are assembled in multimeric structures that constitute the mature functional SP-A although single gene products have been shown to be functional with a varying degree of activity (18, 20, 27, 40–42, 67, 68, 72, 73). SP-A is the most abundant protein of pulmonary surfactant, a lipoprotein complex essential for life. It is an important player of the innate immune system, which constitutes the first line of defense against inhaled challenges. Among its multiple roles, SP-A is an important modulator of lung function, surfactant structure, host defense, and inflammation (17, 57). As a host defense molecule, SP-A regulates a variety of immune cell functions, such as alveolar macrophage cytokine secretion and phagocytosis of bacteria (52, 75).

Several genetic variants have been identified for the SFTPA1 and SFTPA2 genes. These are found in the population with variable frequency, and some of them have been found to be risk factors for lung disease, including tuberculosis, idiopathic pulmonary fibrosis, lung cancer, bronchopulmonary dysplasia, respiratory distress syndrome, and various alveolar and airway infections (15, 24, 51, 55, 58). The most commonly found variants (>1% in the general population) are 6A, 6A², 6A³, and 6A⁴ for SFTPA1, and 1A, 1A⁰, 1A¹, 1A², 1A³, and 1A⁵ for SFTPA2 (Fig. 1). Of these, 6A² and 1A⁰ are found with the highest frequency (13). Moreover, associations of specific SP-A genotypes have been found with varying mRNA expression levels of SP-A (30). Recently, certain SP-A2 variants were associated with low levels of SP-A, and low SP-A levels were associated with poor outcome in patients with lung transplants (6). We have previously hypothesized that the SP-A functional activity depends on the relative amount of SP-A1 and SP-A2, as alterations in both total SP-A content and the SP-A1/SP-A ratio have been observed in lung secretions of patients suffering from asthma and cystic fibrosis (58, 61, 74). Thus it is likely that independent regulatory mechanisms control the expression of variants of SFTPA1 and SFTPA2. In support of this, both splice and sequence differences at the mRNA 5′ and 3′ untranslated regions (UTRs) of SFTPA1 and SFTPA2 have been identified and shown to differentially affect SP-A expression in various experimental models (31, 34, 59, 69, 70). While differences at the 5′UTR usually affect gene transcription and translation (48, 59, 69, 71), sequence variation at the 3′UTR of SP-A variants are shown to primarily affect translation efficiency (60, 70). The impact of this regulation on SP-A1 and SP-A2 protein levels may contribute to individual differences in lung disease susceptibility.

Fig. 1.

Reporter constructs used in this study. The complete 3′ untranslated region (UTR) sequences of SFTPA1 and SFTPA2 variants were cloned downstream of the Firefly luciferase gene (LUC) in our previously generated vector (60). The position and sequence of the variant specific 3′UTR 11-nt indel element is indicated. All vectors contain the SV40 promoter and the 5′UTR sequence variant ABD (31). The corresponding GenBank ID numbers and most frequent variants for each gene are shown.

MicroRNAs (miRNAs) are small noncoding RNAs that control gene expression at the posttranscriptional level by specific interactions with complementary mRNA 3′UTR sequences (9). Numerous studies have reported altered miRNA expression in the development and progression of a variety of human diseases, indicating that miRNAs play a pivotal role in maintaining cell homeostasis and are master regulators of a wide spectrum of biological functions (28, 44, 45, 49, 54, 64). Although the length of most human miRNAs is on average 22 nt, sequence complementarity of only 7 nt at the miRNA “seed region” (positions 2–8 of the miRNA) or even 6 nt (positions 2–7) with the target mRNA is often sufficient to repress translation (36). Depending on the degree of complementarity of the seed region to the matching mRNA sequence, miRNAs can repress gene expression by either translation inhibition and/or mRNA degradation (9, 65). Furthermore, single-nucleotide polymorphisms (SNPs) and polymorphisms at the 3′UTR of genes that alter miRNA binding may affect protein expression associated with the development of various diseases (8, 43, 56).

Our hypothesis is that sequence polymorphisms at the 3′UTRs of SP-A variants can serve as differential binding sites for miRNAs (60). Previously, by using online bioinformatics tools, we have identified candidate sequences for miRNA binding at the 3′UTRs of several SFTPA1 and SFTPA2 variants that may account for differential regulation of SP-A translation (60). In this work, we have investigated whether the candidate miRNA binding sequences play indeed a role in SP-A regulation. Specifically, we studied the most frequently found SFTPA1 and SFTPA2 variants and the effect on gene expression of miRNAs targeting SP-A 3′UTRs, using luciferase reporter assays (46). We found that specific sequence elements of SFTPA1 and SFTPA2 variants differentially affect reporter gene expression. Furthermore, synthetic miRNA mimics had a differential effect on the expression of reporters carrying different SFTPA1 and SFTPA2 sequences, as well as in both SP-A mRNA and protein levels, indicating that sequence polymorphisms at the 3′UTRs of the SP-A genes can affect miRNA binding and translational repression in lung cells.

MATERIALS AND METHODS

SP-A 3′UTR reporter vectors.

Constructs with the luciferase reporter gene upstream of the complete 3′UTR sequence of SFTPA1 variants 6A, 6A², 6A³, and 6A⁴ (1,293–1,304 nt, Fig. 2) and the SFTPA2 variants 1A, 1A⁰, 1A¹, 1A², 1A³, and 1A⁵ (1,309 nt, Fig. 3) were obtained. For this, we modified our previously generated reporter vectors (60) by removal of the T7 promoter and insertion of SV40 and by insertion of previously cloned cDNA sequences of the corresponding 3′UTRs (70). We also generated a control vector without 3′UTR sequence. To avoid any potential effects of 5′UTR varied sequences on the regulation of gene expression (48, 60, 69, 71), all vectors generated and used in the present study contain the same 100 nt 5′UTR corresponding to ABD (31).

Fig. 2.

Sequence alignment of the complete 3′UTR (1293-1304 nt) of SFTPA1 variants 6A, 6A², 6A³, and 6A⁴. mRNA sequences were obtained from GenBank: HQ021433 (6A), HQ021434 (6A²), HQ021435 (6A³), HQ021436 (6A⁴). Sequence alignments were performed with Bioedit (http://www.mbio.ncsu.edu/bioedit/bioedit.html).

Fig. 3.

Sequence alignment of the complete 3′UTR (1309 nt) of SFTPA2 variants 1A, 1A⁰, 1A¹, 1A², 1A³, and 1A⁵. mRNA sequences were obtained from GenBank: HQ021432 (1A), HQ021421 (1A⁰), HQ021422 (1A¹), HQ021423 (1A²), HQ021424 (1A³), HQ021425 (1A⁵). Sequence alignments were performed with Bioedit.

A differential 11-nt deletion/insertion variation element (RefSNP rs368700152, present in all frequently observed SFTPA2 variants and the SFTPA1 variant 6A², but absent in the rest of the most frequently observed SFTPA1 variants) was removed from the corresponding 3′UTR sequence (positions 402–412) by site-directed mutagenesis. The element (CCCACTGCCTG) was also inserted in the SFTPA1 variants 6A, 6A³, and 6A⁴ (Fig. 1). Site-directed mutagenesis reactions were performed with specific primers and the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies, Palo Alto, CA) following the manufacturer's protocol. The identity of all 21 vectors generated was confirmed by DNA sequencing at the Molecular Genetics Core Facility, Pennsylvania State University College of Medicine.

Cell culture.

The lung adenocarcinoma cell line NCI- H441 (ATCC no. HTB-174™, passage 74) was obtained from the American Type Culture Collection (Manassas, VA) and grown in RPMI-1640 medium (Life Technologies, Carlsbad, CA) containing 10% FBS (BenchMark; Gemini Bio Products, Woodland, CA), 1% l-glutamine (Sigma, St. Louis, MO), and 1× antimycotic-antibiotic solution (Sigma). Cells were maintained at 37°C in 5% CO₂ and subcultured weekly.

Three different stably transfected Chinese hamster ovary (CHO-K1) cell lines expressing the human SP-A variants 1A⁰, 6A², and 6A⁴ (generated by us previously, 73) were also used. These were grown in GMEM (Life Technologies, Carlsbad, CA) supplemented with 10% FBS (Benchmark, Gemini Bio Products), 1× antimycotic/antibiotic solution (Sigma), 0.06 mg/ml l-glutamic acid, 0.06 mg/ml l-asparagine, 1× nucleoside mixture (Life Technologies), and 1× sodium pyruvate (Invitrogen, Carlsbad, CA). Cells were cultured at 37°C in 5% CO₂ and subcultured weekly.

miRNA mimics.

The generated vectors and mutants were cotransfected with a Renilla luciferase expression vector for normalization and in the presence or absence of commercial mirVana miRNA mimics (Life Technologies), or a control miRNA mimic (medium GC, Life Technologies). The control mimic is a random sequence that represents a nonspecific miRNA molecule. This mimic has been extensively tested by the manufacturer in human cell lines and tissues and validated to not produce identifiable effects on any known miRNA function. A transfection with the BLOCK-iT fluorescent oligo (Life Technologies) was performed to assess the transfection efficiency of mimics. Block-iT is a conjugated fluorescent small RNA with no significant sequence similarity to human transcripts. Both the control and BLOCK-iT mimics are validated random miRNA sequences that do not produce significant effects on gene expression. MiRNA mimics were purchased for human miRNAs whose seed region was predicted by TargetScan (19) to specifically interact with the surrounding region of the 11-nt element (60). The following eight mimics were used in all experiments: hsa-mir-183-5p, hsa-mir-449b-5p, hsa-mir-612, hsa-mir-762, hsa-mir-767-3p, hsa-mir-3940-5p, hsa-mir-4417, and hsa-mir-4507.

Cell transfection for luciferase assays.

Transfections of vectors and cotransfections of vectors/miRNA mimics were performed on NCI-H441 cells plated in 24-well plates at 30–50% confluence using Lipofectamine 2000 (Life Technologies), following the company's protocol. Both control and miRNA mimics were used at 30 nM final concentration, and 80 ng of reporter vectors were transfected per well. Cells were harvested 48 h after transfection using Passive Lysis Buffer 1× (Promega, Madison, WI), and luciferase activity was monitored in a 20-μl aliquot using the Dual-Luciferase reporter assay system (Promega). Firefly and Renilla luciferase activity were recorded with a FB12 luminometer (Zylux, Maryville, TN) and expressed as the ratio of Firefly Luciferase activity to Renilla luciferase activity.

Cell transfection for gene expression analysis.

For mRNA levels, both NCI-H441 and CHO-K1 cells were plated in 24-well plates and transfected with 30 mM miRNA mimics using Lipofectamine RNAiMAX (Life Technologies), or the Block-iT Fluorescent oligo (10 nM) (Life Technologies). Transfection of oligos into cells was monitored by fluorescence microscopy at 24 and 48 h (Fig. 4). Cells were harvested 48 h after transfection by addition of Trizol (Life Technologies). Total RNA was purified with the Direct-zol RNA-MiniPrep-kit (Zymo Research, Irvine, CA), and RNA concentration and quality were confirmed by Nanodrop and Bioanalyzer. The expression of SFTPA1, SFTPA2, and 18S mRNAs was detected by real-time PCR with specific TaqMan assays (Life Technologies). Results were monitored and analyzed with the QuantStudio 12K Flex Real-Time PCR System (Life Technologies) at the Penn State College of Medicine Genomics Sciences Core Facility.

Fig. 4.

Cell transfection. NCI-H441 cells were cultured and transfected with the control mimic (left) or with a fluorescent oligonucleotide (Block-IT, right). Fluorescence microscopy confirmed intracellular localization of oligos both at 24 and 48 h posttransfection.

Cell transfection for Western blot analysis.

For protein expression analysis, stably transfected CHO-K1 cells expressing SP-A variants (1A⁰, 6A², and 6A⁴) were plated in 12-well plates and transfected with 30 mM of the described miRNA mimics using Lipofectamine RNAiMAX (Life Technologies) for 48 h. Cells were harvested by addition of RIPA buffer (Thermo, Rockford, IL) and protein extraction. Protein concentration was determined by BCA assay (Thermo). Western blots were performed for total SP-A using an antibody that recognizes both SP-A1 and SP-A2 proteins (61). GAPDH (Millipore, Billerica, MA) was used as a loading control.

Antagomir-183-5p transfection.

NCI-H441 cells were plated in 24-well plates and transfected with 30 mM of mimic hsa-mir-183-5p, antagomir-183-5p, or a combination of both with Lipofectamine RNAiMAX, as noted above. Cells were harvested at 48 h posttransfection, and expression of SP-A1 and SP-A2 was determined by real-time PCR.

Statistical analysis.

At least four independent transfection experiments were carried out for each vector, mimic, and vector-mimic combination. Luciferase activity was monitored in triplicate and normalized to the respective controls. SFTPA1 and SFTPA2 mRNA levels were normalized to 18S. Protein expression was analyzed by densitometry and normalized to GAPDH. Statistical analyses were performed using the standard program software SigmaStat version 3.5. Differences among groups were assessed by the ANOVA test or multiple-comparison ANOVA (Tukey's test). Results are expressed as means ± SE. Statistically significant differences were considered when P < 0.05.

RESULTS

Sequence polymorphisms exist at the SFTPA1 and SFTPA2 3′UTRs.

To identify potential differential regulatory regions in the 3′UTRs of SP-A variants, we performed sequence alignments of the variants most frequently found in the population (Fig. 1) by using the Bioedit software (Ibis Biosciences, Carlsbad, CA). Sequences were highly similar among SFTPA1 and SFTPA2, as previously described (60, 70), and several SNPs were identified throughout the sequences of SFTPA1 variants (Fig. 2) and SFTPA2 variants (Fig. 3). As reported previously (26, 53, 60, 70), an insertion/deletion (indel) variable element of 11 nt (CCCACTGCCTG) is found at position 402 of the 3′UTR (60, 70). This element corresponds to the NCBI Reference SNP (refSNP) Cluster Report rs368700152 and is present in all variants from the SFTPA2 gene and the SFTPA1 variant 6A². The indel element has been previously shown to affect translation efficiency (60).

SFTPA1 and SFTPA2 3′UTRs affected reporter expression with variable efficiency.

Reporter constructs with 3′UTRs of SFTPA1 and SFTPA2 variants downstream of the luciferase gene and the ABD 5′UTR regulatory sequence (Fig. 1) were transfected onto NCI-H441 cells, a lung adenocarcinoma cell line that expresses SP-A. We have previously used similar approaches to evaluate the impact of SP-A 5′UTR and 3′UTR variants on translation, either by transient transfection of in vitro transcribed mRNAs of SP-A variants (60) or reporter vectors lacking 5′UTRs in response to dexamethasone (70). The current approach evaluates the independent role of all most frequently expressed 3′UTR sequences of SP-A variants in the context of reporter constructs frequently used for miRNA regulation studies (23, 46). Figure 5 shows the Firefly/Renilla luciferase activity ratio 48 h after transfection with the reporter constructs, normalized to the control (no 3′UTR) vector. We found that the presence of an SP-A 3′UTR sequence downstream of luciferase, independently of the variant used, significantly reduced luciferase expression. We also found that, in this model, the SFTPA1 variant 6A² and the SFTPA2 variants 1A and 1A¹ were more efficient in this inhibition than the rest of the variants. Nucleotide differences in the 3′UTRs of these variants may result in altered secondary structure, mRNA stability, and/or binding of miRNA or protein factors that could explain these differences.

Fig. 5.

Differential negative effect of SFTPA1 and SFTPA2 3′UTR sequences on reporter gene activity. Reporter constructs containing the Firefly luciferase gene upstream of 3′UTRs of SFTPA1 and SFTPA2 variants (or no 3′UTR control) were cotransfected in NCI-H441 cells with a control plasmid containing a Renilla luciferase. Results are expressed as the ratio of Firefly/Renilla luciferase activity relative to control (no 3′UTR) 48 h after transfection. All 3′UTRs decreased gene expression compared with control (*P < 0.001). Variants 6A², 1A, and 1A¹ exhibited the lowest expression (n = 6, **P < 0.05).

The 11-nt element inhibited the expression of the reporter gene independently of the 3′UTR used.

By site-directed mutagenesis, we deleted (del) or inserted (ins) the 11-nt element from the 3′UTRs of SFTPA1 (6Ains, 6A²del, 6A³ins, and 6A⁴ins) and SFTPA2 variants (1Adel, 1A⁰del, 1A¹del, 1A²del, 1A³del, 1A⁵del). Luciferase activity measured 48 h after transfection of NCI-H441 cells with these vectors is shown in Fig. 6. Results were normalized to the corresponding naturally occurring sequence (e.g., 6A vs. 6Ains, 1A vs. 1Adel, etc.). We found that reporter gene activity was significantly decreased when the 11-nt element was inserted at position 402 of variants 6A, 6A³, and 6A⁴. The opposite effect was found when this sequence element was deleted from the variants where it naturally occurs, i.e., 6A², and all SFTPA2 variants (1A, 1A⁰, 1A1, 1A², 1A³, and 1A⁵). Together, these results indicate that the SP-A 11-nt element has a negative effect on gene expression under normal conditions and are consistent with our previous work where variants 6A², 6A³, 6A⁴, and their corresponding mutants were studied for their impact of this element on translation and mRNA stability (60).

Fig. 6.

Effect of removal (del) or insertion (ins) of the 3′UTR 11-nt element on reporter gene activity. Reporter constructs containing the Firefly luciferase gene upstream of the SFTPA1 or SFTPA2 3′UTRs as well as deletion/insertion mutants of the 11-nt indel element were cotransfected in NCI-H441 cells with a Renilla luciferase control plasmid. Results are expressed as the ratio of Firefly/Renilla luciferase activity relative to the respective control (wild-type 3′UTR). Insertion of the 11-nt sequence decreased gene expression in all cases. Deletion of the 11-nt element resulted in increased luciferase activity 48 h posttransfection (n = 6, *P < 0.05).

TargetScan predictions identified several miRNA binding sites at the 11-nt surrounding sequence.

We have previously identified a number of miRNA binding sites at the 3′UTR of SFTPA1 and SFTPA2 variants, as well as within the surrounding sequence of the 11-nt element, using the online tool PITA, which performs predictions on manually entered 3′UTRs and miRBase v.11 (60). Here, we expanded our analysis by using the online tool TargetScan v6.2 (www.targetscan.org) to predict binding of human miRNAs to the SFTPA1 and SFTPA2 3′UTRs annotated in GenBank and using the most updated miRBase version (v.20). Although we found similar results using both approaches, the current approach identified more miRNA targets. We selected eight miRNAs whose seed regions matched different sequence elements among SP-A variants between nucleotides 380–440 of the 3′UTR. These are as follows: hsa-mir-183-5p, hsa-mir-449b-5p, hsa-mir-612, hsa-mir-762, hsa-mir-767-3p, hsa-mir-3940-5p, hsa-mir-4417, and hsa-mir-4507 (Fig. 7). We obtained the mature miRNA sequence of these from miRBase (release 20) (www.mirbase.org) and predicted miRNA binding to SP-A variants by seed-sequence match comparisons. We found that miRNA mir-183-5p is predicted to bind all SFTPA1 and SFTPA2 variants, whereas the rest of the miRNAs selected were predicted only to bind specific variants (Fig. 7). We tested these experimentally by cotransfection with reporter vectors in NCI-H441 cells.

Fig. 7.

Sequence alignment of SFTPA1 and SFTPA2 variants at the 11-nt element region and miRNA binding predictions. A: Bioedit alignment of sequences used in this study for SFTPA1 (top) and SFTPA2 (bottom). The 11-nt element as well SNPs among variants are indicated by *. B: sequence, identity, and predicted binding sites of miRNAs used in this study. The seed region is indicated in bold for the mature miRNA sequence and its reverse complement. TargetScan predictions “7mer-m8” indicate exact match to positions 2–8 of the mature miRNA seed sequence, and “7mer-1A” indicates exact match to positions 2–7 of the mature miRNA followed by the nucleotide A.

Effect of miRNA mimics on SFTPA1 and SFTPA2 3′UTRs reporter vector expression.

To study the effects of the predicted miRNAs on SP-A regulation, we cotransfected reporter vectors of the SP-A 3′UTRs with mimics of the human miRNA predicted by TargetScan (Fig. 7) in NCI-H441 cells. Efficient transfection of oligos into cells was verified by using a fluorescent oligo (Block-iT) and by monitoring intracellular fluorescence at 24 and 48 h posttransfection (Fig. 4). Table 1 shows a summary of the effects of cotransfection with miRNA mimics on luciferase reporter activity for SFTPA1 and SFTPA2 variants 48 h posttransfection. We found that 1) the expression of variants with the 11-nt element (SFTPA1 variant 6A², and all SFTPA2 variants) was affected by the same miRNA mimics, 2) two miRNA mimics (hsa-mir-767-3p and hsa-mir-4507) negatively affected reporter expression for SFTPA1 variants lacking the 11-nt element (6A, 6A³, and 6A⁴), and 3) recombinant reporter vectors containing 3′UTRs of either gene were affected by cotransfection with mimic hsa-mir-767-3p.

Table 1.

Effect of miRNA mimics on SFTPA1 and SFTPA2 reporter vector expression

miRNA Mimic	SFTPA1 Variants				SFTPA2 Variants
	6A	6A2	6A3	6A4	1A	1A0	1A1	1A2	1A3	1A5
hsa-mir-183-5p		↓			↓	↓	↓	↓	↓	↓
hsa-mir-449b-5p		↓			↓	↓	↓	↓	↓	↓
hsa-mir-612		↓			↓	↓	↓	↓	↓	↓
hsa-mir-762
hsa-mir-767-3p	↓	↓	↓	↓	↓	↓	↓	↓	↓	↓
hsa-mir-3940-5p
hsa-mir-4417
hsa-mir-4507	↓		↓	↓

miRNAs 183, 449b, and 612 inhibited expression of variants containing the 11-nt element.

Cotransfection of miRNA mimics for the human miRNAs hsa-mir-183-5p, hsa-mir-449b-5p, and hsa-mir-612 with vectors containing the SFTPA1 variant 6A² (Fig. 8) and all SFTPA2 variants (Fig. 9) resulted in significantly lower reporter expression compared with the control mimic (indicated as C). This effect was abolished by removal of the 11-nt element in all cases, and it was not observed in variants that lacked the element (Figs. 8 and 9). These results indicate that the rs368700152 (11 nt) polymorphism may differentially affect miRNA regulation of SP-A variants. However, no effect was observed when vectors were cotransfected with miRNA mimics for hsa-mir-762, hsa-mir-3940, and hsa-mir-4417, also predicted to bind the 11-nt element by complementary seed sequence. These results show that seed-sequence complementarity is not sufficient for miRNA function, as previously reported for other genes (23, 32).

Fig. 8.

Effect of miRNA mimics on reporter expression vectors of SFTPA1 3′UTRs. Reporter vectors with the Firefly luciferase gene upstream of the SFTPA1 variants 6A, 6A², 6A³, and 6A⁴, as well as mutants of the 11-nt element 6Ains, 6A²del, 6A³ins, and 6A⁴ins were cotransfected in NCI-H441 cells with miRNA mimics and a Renilla luciferase control plasmid. Luciferase activity was measured 48 h after transfection and is indicated as the relative Firefly/Renilla activity normalized to the control mimic (n = 6). MiRNA mimics hsa-mir-183-5p (183), hsa-mir-449b-5p (449b), and hsa-mir-612 (612) significantly decreased luciferase expression vs. the control mimic (C) in vectors with the 11-nt element (6A², 6Ains, 6A³ins, and 6A4ins). Mimic hsa-mir-4507 (4507) affected only 6A, 6A³, and 6A⁴ variants, which naturally lacked the 11-nt element, but not 6A²del, from which the 11-nt element was removed. Hsa-mir-4507 also affected expression in mutants 6Ains, 6A³ins, and 6A⁴ins (see discussion). Mimic hsa-mir-767-3p (767) affected all variants. No effect was observed for the rest of the mimics. *P < 0.05.

Fig. 9.

Effect of miRNA mimics on reporter expression vectors of SFTPA2 3′UTRs. Reporter vectors with the Firefly luciferase gene upstream of the SFTPA2 variants 1A, 1A⁰, 1A¹, 1A², 1A³, and 1A⁵ and deletion mutants of the 11-nt element (del) were cotransfected in NCI-H441 cells with miRNA mimics and the Renilla luciferase control plasmid. Luciferase activity was measured 48 h after transfection and is indicated as the relative Firefly/Renilla activity normalized to the control mimic (n = 6). Three miRNA mimics, hsa-mir-183-5p (183), hsa-mir-449b-5p (449b), and hsa-mir-612 (612), significantly decreased luciferase expression of vectors with the 11-nt element vs. control mimic (C). No effect of these miRNA mimics was observed when the 11-nt sequence was removed from these 3′UTRs. Mimic hsa-mir-767-3p (767) affected all naturally occurring variants and mutants. No effect was observed for the rest of the mimics studied. *P < 0.05.

miRNA 4507 is a selective inhibitor of SFTPA1 variants 6A, 6A³, and 6A⁴.

When we cotransfected miRNA mimic hsa-mir-4507 with recombinant reporter vectors containing 3′UTRs of SFTPA2 variants and mutants of the 11-nt element, we found that this miRNA mimic selectively affected expression of constructs with the 3′UTRs of SFTPA1 variants 6A, 6A³, and 6A⁴ (Figs. 8 and 9). In this case, insertion of the 11-nt element did not affect expression levels, as the predicted binding site of this miRNA is located upstream of the indel element region. These results indicate that a mismatch in position 5 of the seed region of miRNA mir-4507 (as shown in Fig. 7) was sufficient to prevent miRNA binding to the SFTPA2 variants and the SFTPA1 variant 6A². This mismatch is a SNP (G/A) at position 401 of the 3′UTR, with 6A² and the SFTPA2 variants having a G at this position.

miRNA 767 inhibited expression of reporters with both SFTPA1 and SFTPA2 variants.

The expression of reporter constructs with all variants, as well as their corresponding mutants, was inhibited by cotransfection with mimic hsa-mir-767-3p (Figs. 8 and 9). Although this miRNA was predicted to bind all SFTPA2 variants and only the 6A² SFTPA1 variant (Fig. 8), a SNP (A/G) at position 435 in variants 6A, 6A³, and 6A⁴ (Fig. 8) was not sufficient to prevent miRNA function, indicating that Watson-Crick matching of only 6 nt at the seed region can also allow miRNA repression, as demonstrated previously in other systems (2).

Together, these results indicate that specific interactions at the 11-nt element of SP-A variants may affect miRNA regulation of gene expression. Cotransfection of miRNA mimics with mutants of the 11-nt element for each SP-A variant confirmed our findings.

miRNA mimics affect SFTPA1 and SFTPA2 mRNA and protein expression.

To study the effects of miRNAs on SP-A expression, we transfected mimics in a cell line that expresses SP-A (NCI-H441, homozygous for the SFTPA1 variant 6A⁴, and the SFTPA2 variant 1A⁵) (4, 26). At 48 h posttransfection, we found that mir-767 and mir-4507 significantly decreased SFTPA1 mRNA levels compared with cells transfected with control mimic (C), BLOCK-iT (B), or untreated cells (U) (Fig. 10, left). In the case of SFTPA2, transfection with mir-183, mir-449b, mir-612, and mir-767 resulted in significantly lower levels of mRNA (Fig. 10, right).

Fig. 10.

Effect of miRNA mimics on SFTPA1 and SFTPA2 mRNA levels in NCI-H441 cells. NCI-H441 cells (genotype 1A⁵/1A⁵-6A⁴/6A⁴) were transfected with miRNA mimics. Expression of SFTPA1 and SFTPA2 was measured by real-time PCR 48 h after transfection and normalized to 18S. MiRNA mimics hsa-mir-183-5p (183), hsa-mir-449b-5p (449b), and hsa-mir-612 (612) significantly decreased SFTPA2 mRNA levels but not SFTPA1 levels compared with the control mimic. Mimic hsa-mir-4507 (4507) only affected SFTPA1 mRNA. Mimic hsa-mir-767-3p (767) decreased levels of both SFTPA1 and SFTPA2 mRNAs (n = 6, *P < 0.05). H₂O, cells transfected with water; U, cells nontransfected; B, cells transfected with Block-iT fluorescent oligo.

NCI-H441 cells express both SP-A1 and SP-A2 proteins. Although specific antibodies for SP-A1 (61) and SP-A2 (commercially available) exist, these differ in sensitivity. Thus, to analyze the effects of mimic in protein expression, we used stably transfected CHO-K1-SP-A cells that express single SP-A variants, generated by us previously (73). This approach allowed the study of the effects of miRNA transfection on each gene variant separately. We used CHO-K1-SP-A cells for two SP-A1 variants (6A² and 6A⁴) and one SP-A2 variant (1A⁰). Transfection of oligonucleotides into cells was verified using the BLOCK-iT oligo (data not shown). We also studied the effect of mimic transfection on SFTPA1 and SFTPA2 mRNA and protein levels (Fig. 11). We found that mir-767 significantly decreased mRNA and protein levels of both SFTPA1 6A² and 6A⁴ variants, as well as the SFTPA1 variant 1A⁰. In addition, mir-183, mir-449b, and mir-612 significantly decreased SFTPA1 and SFTPA2 mRNA and protein levels in CHO-K1 cells expressing the SFTPA1 variant 6A² and the SFTPA2 variant 1A⁰, both containing the 11-nt element. Consistent with luciferase reporter assays, mir-4507 affected the expression of only SP-A1 variant 6A⁴, but not 6A² and 1A⁰ (Fig. 11). Interestingly, two miRNA mimics (mir-4417 and -4507) affected SP-A protein expression but not mRNA levels when transfected in CHO cells expressing the SP-A1 variant 6A².

Fig. 11.

Effect of miRNA mimics on single variant expression in Chinese hamster ovary (CHO)-K1 cells. CHO-K1 stably transfected cells expressing SP-A variants 1A⁰, 6A², and 6A⁴ were transfected with miRNA mimics. Expression of SFTPA1 and SFTPA2 mRNA was measured by real-time PCR 48 h after transfection and normalized to 18S (top). Expression of surfactant protein A (SP-A) protein was detected by Western blot using an antibody that recognizes both SP-A1 and SP-A2 and normalized to GAPDH (bottom). MiRNA mimics hsa-mir-183-5p (183), hsa-mir-449b-5p (449b), and hsa-mir-612 (612) significantly decreased expression of the SP-A2 variant 1A⁰ and the SP-A1 variant 6A², both containing the 11-nt element, but did not affect the expression of the SP-A1 variant 6A⁴, compared with the control mimic. Mimic hsa-mir-4507 (4507) only affected SP-A1 variant 6A⁴ expression. Mimic hsa-mir-767-3p (767) affected expression levels of all variants. Two miRNA mimics (mir-4417 and -4507) affected SP-A1 protein expression but not mRNA levels in CHO-K1 cells expressing the variant 6A² (n = 3, *P < 0.05).

Effect of mir-183 on SFTPA1 and SFTPA2 gene expression.

In a previous work, we have confirmed the expression of hsa-mir-183 in the NCI-H441 cell line (60). To validate the observed specific effects of mir-183 on the expression of SFTPA1 and SFTPA2 variants containing or lacking the 11-nt element, we transfected NCI-H441 cells (genotype 1A⁵/1A⁵-6A⁴/6A⁴) with a synthetic small RNA with perfect complementary sequence to mir-183 (antagomir-183, A183) to antagonize mir-183 effects. Figure 12 shows that, as expected, transfection of A183 into NCI-H441 cells did not affect mRNA levels of SFTPA1, compared with levels in cells transfected with mir-183 or the control mimic. However, SFTPA2 mRNA levels were significantly higher in cells transfected with A183 vs. control mimic. Moreover, transfection with A183 abolished the inhibitory effects of mir-183 on SFTPA2 mRNA levels, suggesting a direct interaction of mir-183 at the 3′UTR 11-nt element of SFTPA1 variant 1A⁰.

Fig. 12.

Transfection of NCI-H441 cells with antagomir-183 reverses the effect of mir-183 on SFTPA2 mRNA levels. NCI-H441 cells (genotype 1A⁵/1A⁵-6A⁴/6A⁴) were transfected with mir-183 mimic, antagomir-183 (A183), or both. Expression of SFTPA1 and SFTPA2 mRNA was measured by real-time PCR 48 h after transfection and normalized to 18S. Transfection with miRNA mimic hsa-mir-183-5p (183) significantly decreased SFTPA2 mRNA levels but did not affect SFTPA1 levels compared with the control mimic. Antagomir-183 abolished this effect when cotransfected with mir-183. Furthermore, transfection with A183 resulted in increased expression levels of SFTPA2 mRNA but had no effect on SFTPA1 levels (n = 3, *P < 0.05).

DISCUSSION

SP-A is a lung innate immunity molecule that plays a role in maintaining normal lung function, host defense, and inflammation. Studies of the genetics of SP-A have revealed correlations of SFTPA1 and SFTPA2 variants with lung disease susceptibility (16, 58), and polymorphisms of SP-A variants have been associated with variable gene expression levels (6). Furthermore, the relative levels of SP-A1 to total SP-A protein have been found to differ in samples from healthy subjects and subjects with lung disease (61, 74). The 3′UTRs of SFTPA1 and SFTPA2 mRNAs show sequence polymorphisms for which variant-specific miRNA-predicted binding sites have been found in silico, further pointing to SP-A gene and variant-specific regulatory differences. The goal of this study was to evaluate the effects of 1) 3′UTR sequences of SFTPA1 and SFTPA2 variants, 2) an 11-nt element (present in all SP-A2 variants but absent in most SP-A1 variants), and 3) miRNAs predicted to bind sequences at 11-nt indel region and surrounding sequences, on the expression of a reporter gene (luciferase), stably transfected cell lines with SP-A variants, or SP-A gene expression in lung cells. We found a differential impact of a number of human miRNAs on the expression of reporter constructs containing 3′UTRs of SFTPA1 and SFTPA2, as well as in the expression of SP-A mRNA and protein levels, indicating that miRNAs can differentially mediate regulation of expression of SP-A variants. The effects of this regulation on SP-A1 and SP-A2 levels may contribute to individual susceptibility to lung disease.

The specific regulatory mechanisms of expression of human SP-A variants are still not well understood. We have recently investigated whether mRNA regulatory elements interact to regulate translation of SFTPA1 and SFTPA2. We found differences in the translation efficiency of in vitro prepared transcripts containing variants of both SP-A 5′UTR and 3′UTRs (48, 59, 60). In one study, we found differences in translation efficiency in vitro for mRNAs containing 3′UTRs of some SP-A variants (60). Variants lacking the 11-nt element showed the highest translation efficiency when these mRNAs were translated in vitro using rabbit reticulocyte lysates or when they were transfected into NCI-H441 cells. Insertion of the 11-nt element in variants that lacked it resulted in reduced translation, and removal of this element resulted in increased translation efficiency (60). In the present work, we have extended our studies to all SP-A variants found in the population with frequency greater than 1% (14), and we have used a 3′UTR reporter construct approach (46), which allows cotransfection with miRNA mimics and controls, and further analysis of miRNA regulation of gene expression. Our findings support previous observations and indicate that, in all cases, the 11-nt element can repress gene expression either when it is naturally occurring at the 3′UTR of SFTPA2 variants and the SFTPA1 variant 6A² or when it is inserted at position 402 by site-directed mutagenesis. Removal of the 11-nt element had the opposite effect. Thus it is likely that specific cellular factors such as the recently discovered miRNAs interact with this regulatory sequence and control protein expression of SP-A1 and SP-A2.

MicroRNAs are small nonprotein-coding RNA molecules that interact with short sequences at the 3′UTR of transcripts to negatively impact translation by either direct repression of the translation process and/or by targeting the mRNA for degradation (50). There are more than 2,500 human miRNAs annotated in the current miRBase, and their function has been studied by the use of numerous tools and experimental approaches. Preliminary bioinformatics data revealed potential binding sites for miRNAs in the 11-nt region of SP-A variants (60). Most of these miRNAs are expressed in lung tissue, and their levels have been shown to be altered under pathological conditions (7, 11, 29, 49, 64, 76). In the present study, we used TargetScan to find miRNAs that would selectively bind to 3′UTRs of SFTPA1, SFTPA2, or both. We selected those whose binding sites were at, or surrounding, the 11-nt element and that were shown to be expressed in human lung cells. Transfection of NCI-H441 cells with mimics for eight human miRNAs revealed a differential impact on the expression of 1) recombinant reporter vectors containing the complete 3′UTR sequences for four human SP-A1 variants and six human SP-A2 variants downstream of luciferase, and 2) reporter vectors containing the 3′UTRs with insertion/deletion of the 11-nt element.

We found selective effects for five of the eight mimics studied. The remaining three mimics (hsa-mir-762, hsa-mir-3940-5p, and hsa-mir-4417) did not show significant effects on luciferase expression although their seed region was complementary to target sequences in the 3′UTRs of the SP-A genes. This is in accordance with previous studies that demonstrated that sequence match at the seed region is necessary, but not sufficient, for miRNA binding to the target mRNA (22). Other factors such as sequence composition near the target site and/or complementarity of the miRNA 3′ end, as well as the spatial structure of the binding site, may account for an efficient miRNA recognition and function (8, 22, 32).

The latter may explain the results observed for hsa-mir-183-5p in the current study. Our results indicate that cotransfection of both SFTPA1 and SFTPA2 3′UTR vectors with a mir-183 mimic significantly reduced reporter gene expression of variants expressing the 11-nt element (Figs. 8 and 9). Furthermore, transfection of NCI-H441 cells (expressing both SFTPA1 and SFTPA2) or CHO-K1 cells expressing single gene variants with mir-183 mimic resulted in a significant reduction of SP-A mRNA and protein levels for variants expressing the 11-nt element (Figs. 10 and 11). Although bioinformatic predictions indicated that this miRNA may bind all SFTPA1 and SFTPA2 variants (Fig. 7), the inhibitory effects for this miRNA were only observed when variants contained the 11-nt element. Although the seed sequence for this miRNA is located at position 417–423 of the 3′UTR (i.e., outside of the indel element), the presence of the 11 nt at position 402–412, either in the variants that naturally contain the indel or when inserted by site-directed mutagenesis, significantly affected the effects of the mir-183-5p mimic (Figs. 8 and 9). As we showed previously, both the secondary structure and the mRNA stability at this region are affected by the presence of the 11-nt element (60).

miRNA 183 belongs to the mir-183/96/182 cluster, known to play important roles in retinal differentiation (39) and in the development of various cancers (10, 37, 78), including lung cancer (78). We have previously identified miRNA expression in NCI-H441 cells as a potential regulator of SP-A translation (60). To further characterize the effects of this miRNA in posttranscriptional regulation of SP-A variants, we transfected NCI-H441 cells with a specific antagomir for mir-183 (35). We found that antagomir-183, not only abolished the effects of transfected mir-183 on SFTPA2 mRNA levels, but also resulted in an increase of mRNA levels of SFTPA2 when transfected alone (Fig. 12). These results indicate that a potential interaction between SFTPA2 mRNAs and the endogenous mir-183 in NCI-H441 cells may result in mRNA degradation under normal conditions. Furthermore, and as expected based on the observed effects on luciferase reporter expression, no effects were observed for SFTPA1 mRNA levels in NCI-H441 cells (genotype 6A⁴/6A⁴) when either mir-183 or the antagomir were transfected.

Two of the miRNAs studied (hsa-mir-449b-5p and hsa-mir-612), whose seed regions were contained within the 11-nt element, showed specific regulatory effects on variants containing the indel element. As expected, the opposite effect was observed when the element was deleted from these variants. Furthermore, both miRNAs inhibited luciferase expression of variants 6A, 6A³, and 6A⁴ only when these were mutated by insertion of the 11 nt. These results indicate that these two miRNAs are likely to differentially regulate expression of SP-A1 and SP-A2 variants that contain the indel element. We confirmed these findings by transfection of mimics in NCI-H441 and CHO-K1 cells expressing SP-A variants (Figs. 10 and 11). In addition, both miRNAs have been shown to be expressed in human lung cells and to regulate a variety of cellular functions including splicing, translation, cell cycle, and cell differentiation (12, 38, 63). Further experiments using antagomirs in human alveolar type II cell cultures are likely to confirm this hypothesis.

In the case of miRNA 4507, predicted to bind at positions 399–405 of the SFTPA1 variants 6A, 6A³, and 6A⁴, a negative effect on luciferase expression was observed for these three variants. As seen in Fig. 7, the first four nt of the 11-nt element (CCCA) are repeated at position 413, where the element ends; thus this effect was observed even when the 11-nt element was inserted (6Ains, 6A³ins, 6A⁴ins, Fig. 8). However, a SNP at nucleotide 401 (G/A) in the 6A² variant and the SFTPA2 variants that caused a mismatch with the seed region of mir-4507 was sufficient to abolish the negative effect of this miRNA on the expression of the reporter gene in all these variants containing the 11-nt element, plus in the mutant 6A²del. Similar results were obtained for the SP-A mRNA levels in NCI-H441 cells, and CHO cells expressing 1A⁰ and 6A² (Figs. 10 and 11). Interestingly, this miRNA mimic was able to decrease protein expression when transfected in 6A² CHO cells even though there was no change in mRNA levels (Fig. 11). These results indicate that imperfect match at the seed region is sufficient to affect protein translation in this cell model, whereas near-perfect to perfect match is needed for mRNA degradation, as observed for other miRNA-mRNA interactions (28, 66). A similar scenario was observed for mir-4417, which appeared to specifically affect protein translation of SP-A1 6A² variant in CHO cells without affecting mRNA levels. Additional experiments using antagomirs in CHO cells could reveal the molecular mechanisms involved in the regulatory differences among variants for mir-4417 and 4507.

Conserved Watson-Crick pairing of the target mRNA to the 5′ region of the miRNA at nucleotides 2–7 (known as the seed region) is the most common mechanism for miRNA target recognition, often called 7-nt canonical pairing. However, several reports have described seed match of only 6 nt or even nonperfect match of the seed region as sufficient for translational repression (2, 3, 47). Here, we have found that, in the case of hsa-mir-767-3p, predicted by TargetScan to bind at position 429–435 of all SFTPA2 variants and the 6A² variant, a SNP causing a change from G to A at position 435 in the SFTPA1 variants 6A, 6A³, and 6A⁴ did not result in significant changes in the ability of this miRNA to repress luciferase expression. This is in accordance with computational reports that established that target sites with a mismatch or a G:U wobble at the seed region of the miRNA-mRNA interaction can still confer a regulatory effect of miRNAs on its target (3). On the other hand, we did not find any effects on luciferase activity for miRNA hsa-mir-762, for which TargetScan predicted a 6-nt matching sequence (CAGCCC) at position 399–405 of variants 6A², 1A, 1A⁰, 1A¹, 1A³, and 1A⁵ (Fig. 7), indicating that additional factors, such as those mentioned above, may affect binding specificity and allow miRNA function at the target site.

A growing number of studies are highlighting the impact of mutations and polymorphisms at miRNA binding sites in gene expression and in the development of a number of diseases (21, 56). These include Tourette's syndrome, Parkinson's disease, metabolic syndrome, and diseases of the lung such as asthma, lung cancer, and cystic fibrosis (1, 33, 62, 77). Moreover, several inflammatory lung diseases have been associated with SFTPA1 and SFTPA2 variants, as well as with various intragenic haplotypes and SNPs in the coding region of both genes (reviewed in 58). Thus it is likely that SNPs and polymorphisms at various miRNA binding sites in the 3′UTR of SP-A variants may affect posttranscriptional gene regulation and contribute to lung disease susceptibility by affecting the SP-A1 to SP-A2 expression ratio (61, 74).

In summary, we have identified several miRNA targets for SFTPA1 and SFTPA2 variants. By using reporter assays in human lung cells, we have experimentally confirmed regulation of gene expression by specific miRNAs targeting sequence polymorphisms at the mRNAs of SP-A variants and the role of a differential sequence element (rs368700152) at position 402 of the SP-A 3′UTRs. We found three miRNAs (mir-183-5p, mir-449b-5p, and mir-612), with a regulatory effect on variants containing the 11-nt element, and we identified one miRNA (mir-4507) that specifically targeted SFTPA1 variants without the element. We confirmed the specific interaction of miR-183 (either endogenous or transfected) and SP-A 3′UTRs in NCI-H441 cells by the use of a specific antagomir. We found that mir-4417 and mir-4507 inhibit translation without affecting mRNA levels of the 6A² variant in CHO cells. Finally, one miRNA (mir-767-3p) was found to affect the mRNA and protein levels of both SFTPA1 and SFTPA2 genes. Together, these results demonstrate that sequence polymorphisms at the 3′UTR of SP-A genetic variants can affect miRNA regulation of gene expression. As growing evidence demonstrates that miRNA levels can vary under different physiological conditions, we postulate that this differential regulation results in variable levels of SP-A1 and SP-A2 in different individuals. Thus, by mediating the differential regulation of translation of SP-A1 and SP-A2 variants, miRNAs may contribute to differences in lung disease susceptibility among individuals, as well as in the response to diverse physiological situations and environmental challenges.

GRANTS

This work was supported by grants from NIH (HL-34788, J. Floros), Children's Miracle Network (P. Silveyra), and Sigma Delta Epsilon-Graduate Women in Science (Adele Lewis Grant Fellowship, P. Silveyra). Dr. Silveyra's research is supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under BIRCWH award number K12HD055882, “Career Development Program in Women's Health Research at Penn State.” The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: P.S. and J.F. conception and design of research; P.S. and S.L.D. performed experiments; P.S. analyzed data; P.S., S.L.D., and J.F. interpreted results of experiments; P.S. prepared figures; P.S. drafted manuscript; P.S., S.L.D., and J.F. edited and revised manuscript; P.S., S.L.D., and J.F. approved final version of manuscript.