A new splice variant of the major subunit of human asialoglycoprotein receptor encodes a secreted form in hepatocytes

Abstract

Background: The human asialoglycoprotein receptor (ASGPR) is composed of two polypeptides, designated H1 and H2. While variants of H2 have been known for decades, the existence of H1 variants has never been reported.

Principal findings: We identified two splice variants of ASGPR H1 transcripts, designated H1a and H1b, in human liver tissues and hepatoma cells. Molecular cloning of ASGPR H1 variants revealed that they differ by a 117 nucleotide segment corresponding to exon 2 in the ASGPR genomic sequence. Thus, ASGPR variant H1b transcript encodes a protein lacking the transmembrane domain. Using an H1b-specific antibody, H1b protein and a functional soluble ASGPR (sASGPR) composed of H1b and H2 in human sera and in hepatoma cell culture supernatant were identified. The expression of ASGPR H1a and H1b in Hela cells demonstrated the different cellular loctions of H1a and H1b proteins at cellular membranes and in intracellular compartments, respectively. In vitro binding assays using fluorescence-labeled sASGPR or the substract ASOR revealed that the presence of sASGPR reduced the binding of ASOR to cells. However, ASOR itself was able to enhance the binding of sASGPR to cells expressing membrane-bound ASGPR. Further, H1b expression is reduced in liver tissues from patients with viral hepatitis.

Conclusions: We conclude that two naturally occurring ASGPR H1 splice variants are produced in human hepatocytes. A hetero-oligomeric complex sASGPR consists of the secreted form of H1 and H2 and may bind to free substrates in circulation and carry them to liver tissue for uptake by ASGPR-expressing hepatocytes.

Figure 1. Sequence analysis of splice variants of ASGPR H1 and identification of polyclonal antibody that can recognize ASGPR H1b specifically.

(A) Amplification of ASGPR H1a and H1b cDNAs from human liver tissue. Lane 1: DNA ladder. Lane 2: RT-PCR products of total RNA from human liver tissue. (B) Sequence comparison of ASGPR H1a and H1b transcripts. Upper part: the relevant parts of ASGPR H1 genomic and mRNA sequences are indicated. The structure of the genomic sequence encoding ASGPR H1 with relevant exons was shown. The 117 bp deletion in ASGPR H1b mRNA sequence corresponds exactly to the second exon (gray box) of the genomic sequence of ASGPR H1. The primers P1 to P6 used for ASGPR H1 cDNA cloning and for detection of the variant sequences are marked at corresponding position. Lower part: the sequence alignment of the genomic sequence of ASGPR H1 and cDNA sequences of H1a and H1b shows the boundaries of the alternatively spliced 117 nt region in the genomic sequence and the consensus splice donor/acceptor sites (underlined). The difference between the mRNA sequences of ASGPR H1a and H1b is boxed. Identical nucleotides are indicated by dots; missing nucleotides are indicated by dashes. (C) Specificity analysis of polyclonal antibodies by western blot. Cell lysates of H1a-HA (lane 1, 3, 5, 7) or H1b-HA (lane 2, 4, 6, 8) transfected cells were separated by SDS-PAGE. Primary antibodies are used as follows: lanes 1 and 2: anti-HA; lanes 3 and 4: anti-H1; lanes 5 and 6: anti-H1b; lanes 7 and 8: non-immunized mouse serum. The positions of H1a-HA and H1b-HA are indicated by arrows.

Figure 2. Localization of ASGPR H1b protein in human liver tissue by immunohistochemistry.

(A) Paraffin sections of human liver tissue were stained with anti-H1 antibody. Staining of both the surface of the cellular membrane (red arrow) and the cytoplasm are observed. (B) Paraffin sections of human liver tissue were stained with anti-H1b antibody. Only cytoplasmic staining is observed. (C) Negative control, stained with sera from non-immunized mice.

Figure 3. Localization of ASGPR H1b protein in transfected cells.

(A) Fusion proteins of the two variants of ASGPR H1 subunit were designed to trace their cellular locations. Both ASGPR H1a and H1b were fused to EGFP in the vector pXF-1E. In pXF-3H-H1a, ASGPR H1a was expressed with an N-terminal HA tag. (B) Transient expression of EGFP-ASGPR H1a in HeLa cells by transfection with the plasmid pXF1E-H1a. Fluorescence signal of both the surface of the cellular membrane and the cytoplasm was observed. (C) Transient expression of EGFP-ASGPR H1b in HeLa cells by transfection with the plasmid pXF1E-H1b. The EGFP-H1b protein is only located in the cytoplasmic compartment. (D) Transient co-expression of EGFP-ASGPR H1a and H1b in HeLa cells by transfection with plasmids pXF3H-H1a and pXF1E-H1b. After co-transfection with both plasmids, HeLa cells were stained with anti-HA antibody and corresponding PE-labeled second antibody to detect ASGPR H1a.

Figure 4. Analysis of functional sASGPR in serum and supernatant of HepG2 cells by SDS-PAGE and western blot.

(A) SDS-PAGE. Lane 1: protein marker. Lane 2: functional sASGPR purified from normal human serum. Lane 3: functional sASGPR purified from supernatant of HepG2 cells. The positions of the fractions are indicated by arrows. (B) Western blot. Functional sASGPR purified from normal human serum (Lanes 1–3) and supernatant of HepG2 cells (Lanes 4–6) were separated by SDS-PAGE. Primary antibodies used are as follows: lanes 1 and 4: anti-H1; lanes 2 and 5: anti-H1b; lanes 3 and 6: anti-H2.

Figure 5. Ligands binding assay.

The binding of fluorescence labeled A-ASOR (A) and A-sASGPR (B) to cells was determined by FACS. (A) Huh7 cells and cell line 4-1-6 were incubated with A-ASOR alone, or with A-ASOR in the presence of s-ASGPR or BSA. Binding of A-ASOR were determined by MFI. HeLa cells were used as a negtive control. The MFI value of the sample with A-ASOR alone was set as 100%. (B) HeLa cells, Huh7 cells, and the ASGPR-expressing cell line 4-1-6 were incubated with A-sASGPR alone, or with A-sASGPR in the presence of ASOR or BSA. Binding of A-sASGPR was determined by MFI. The MFI value of the sample with A-sASGPR alone was set as 100%.

Figure 6. ASGPR H1b expression in hepatoma cells with and without viral replication.

The expression levels of ASGPR H1a (A) and H1b (B) transcripts were determined by real-time PCR. The mRNA levels of ASGPR H1 variants in HepG2 cells HepG2.2.15 with HBV replication, Huh7.5.1 cells, and HCV infected Huh7.5.1 cells were determined. The expression levels of ASGPR H1 in HepG2 and Huh7.5.1 cells were set as 100%. (C) The ratio of H1a/H1b expressed in hepatoma cells.

Figure 7. Possible physiological function of sASGPR.

This diagram depicts the dynamic metabolic turnover of Gal/GalNAc glycoconjugates and their removal by the liver, as described by the Gal/GalNAc homeostasis hypothesis. Various tissues synthesize and secrete different Gal/GalNAc-specific lectins (black half circles) and Gal/GalNAc-containing glycoconjugates (red or blue full circles). The maintenance of these cell-cell or cell-matrix interactions requires the continuous participation of these molecules. Normal tissue turnover, injury, diet, disease, or other events may cause Gal/GalNAc glycoconjugates to be released into the blood. These soluble Gal/GalNAc glycoconjugates will be recognized and bound by sASGPR until they are transported safely to the liver, where they will be removed and degraded by the hepatic ASGPR system (molecules undergoing intracellular degradation are indicated by yellow full circles). Without the help of sASGPR, soluble Gal/GalNAc glycoconjugates will compete for the endogenous cell-cell or cell-matrix interactions between one or more Gal/GalNAc glycoconjugates and their lectin receptors before they can be removed by the hepatic ASGPR system. This competition could result in the disruption of normal tissue organization, cell growth regulation, or other functions necessary for normal healthy tissue homeostasis.