Identification of Two Distinct Cell Binding Sequences in the Vitamin D Binding Protein

Abstract

The vitamin D binding protein (DBP) is a multifunctional, albumin-like plasma protein that often requires cell surface binding to mediate some of its diverse functions. DBP binds to several different molecules on the external face of the plasma membrane indicating that it may possess distinct cell binding sequences. In this report, surface plasmon resonance was utilized to evaluate the relative binding of the human myeloid cell line U937 to immobilized recombinant expressed DBP in order to identify cell localization sequences. U937 cells showed robust binding to immobilized native DBP, but essentially no interaction when sensor chips were coated with β₂-microglobulin or BSA. The cell-DBP interaction was completely eliminated if cells were pretreated with soluble DBP. Recombinant DBP domains and truncated domains were next evaluated to determine the location of cell binding regions. Domains I (amino acids 1–191) and III (379–458), but not domain II (192–378), could support cell binding. Further evaluation of domain I, using truncated proteins and overlapping peptides, demonstrated that a single amino acid sequence, residues 150–172 (NYGQAPLSLLVSYTKSYLSMVGS), mediated cell binding. The domain III cell binding region was investigated using truncated versions of domain III fused to full-length domain II that served as a scaffold. These experiments indicated that the cell binding sequence is located in the first portion of that domain (379–402: ELSSFIDKGQELCADYSENTFTEY). Overlapping peptides spanning this sequence could partially block cell binding only when used in combination. We conclude that DBP contains two cell localization sequences that may be required for some of the multiple functions of this protein.

1. INTRODUCTION

Vitamin D binding protein (DBP) is a multifunctional and highly polymorphic plasma protein synthesized primarily in the liver [1, 2]. DBP (also referred to as Gc-globulin) is a member of the albumin gene family and has the characteristic multiple disulfide-bonded, triple domain structure [1, 2]. The plasma concentration of DBP is 6–7 µM but the protein is ubiquitous in vivo and significant quantities (0.1–1 µM) have been detected in all fluid compartments (cerebrospinal, bronchoalveolar, synovial, etc.) [1, 3]. In contrast to albumin, DBP levels in blood rise modestly (20–50%) during the acute phase inflammatory response [1, 2]. DBP has several distinct and outwardly unrelated functions including a vitamin D transport protein, an extracellular scavenger for G-actin released from necrotic cells, a chemotactic cofactor for the complement activation peptide C5a, and a macrophage and osteoclast activating factor [1, 2]. Ligand binding regions within the 458 amino acid sequence of DBP have been identified: a vitamin D sterol binding segment in the N-terminal domain (amino acids 35 to 49) and a G-actin binding region in the C-terminal domain (amino acids 373 to 403) [4, 5]. More recent work on the crystal structure of DBP (bound to either vitamin D₃ or actin) has confirmed the vitamin D sterol binding site, but has demonstrated that actin interacts with distinct amino acid sequences in all three DBP domains [6–9]. These studies also revealed that DBP is a broad U-shaped or saddle-shaped molecule with domains I and III forming the front and back of the saddle and domain II the seat [6–9]. This shape is designed to perfectly fit a G-actin molecule. Vitamin D sterol binding pocket is distinct in domain I and DBP can bind both ligands independently [4]. Recently, we have identified a 20 amino acid sequence in the N-terminal domain (amino acids 130–149) that is essential for the protein to function as a chemotactic cofactor for C5a [10].

All of the cellular functions of DBP require that the protein binds to its target cell surface. Numerous investigators have reported a cell-associated form of DBP in many cell types including all leukocytes [11–22]. Cell-associated DBP is not a novel cellular form but rather plasma-derived DBP bound to the cell surface [23]. DBP appears to bind with low avidity to multiple cell surface ligands such as chondroitin sulfate proteoglycans [24], megalin [25, 26], cubulin [26], CD44 and annexin A2 [27]. Since the interaction of DBP with plasma membrane ligands is central to it’s cell-mediated functions, the objective of this study was to identify cell binding sequences within DBP. Our experimental approach utilized surface plasmon resonance (SPR) to measure cell binding to DBP in real time using unmodified proteins and peptides. SPR is considerably more sensitive and reproducible than previous methods we have used to study cell binding where DBP was covalently coupled to either ¹²⁵I [24] or Alexafluor-488 [27]. Although SPR usually is employed to quantitate the interaction between two purified molecules, it also can be used to measure the relative binding of cells (suspended in media) to an immobilized ligand [28]. The human myeloid cell line U937 was utilized since these cells grow in suspension and previous data has shown that DBP binding to U937 cells essentially is identical to neutrophils obtained from peripheral blood [29, 30]. In addition, we have recently demonstrated that neutrophil binding to immobilized DBP using SPR [31] is almost identical to that of U937 cells reported herein. Results show that molecules on the surface of U937 cells bind two distinct amino acid sequences in DBP, one in the N-terminal domain (domain I) and the other in the C-terminal domain (domain III). Analysis of the three-dimensional structure of DBP reveal that the location of these sequences, conceivably, could permit DBP to function as an adaptor protein and bridge two distinct cell surface molecules. This multi-ligand binding may be necessary for DBP to mediate its cellular functions.

2. MATERIALS AND METHODS

2.1 Reagents

Human DBP was purified from plasma and obtained from Athens Research and Technology (Athens, GA). Full-length human DBP cDNA (Gc-2 allele, GenBank Accession Number P02774), clone number CS0DM004YF02, was purchased from Invitrogen (Carlsbad, CA). DNA restriction and modification enzymes were purchased from New England Biolabs (Beverley, MA). Oligonucleotides were synthesized by Invitrogen. pGEX-4T-2 expression vector was purchased from Amersham BioSciences (Piscataway, NJ). The IgG fraction of polyclonal goat anti-human DBP was purchased from DiaSorin, (Stillwater, MN) and then affinity purified using immobilized DBP. Protease inhibitor cocktail mixture was obtained from Sigma-Aldrich (St. Louis, MO).

2.2 Cells and cell culture

U937 cells were originally obtained from the American Type Culture Collection (Manassas, VA) and transfected with either the human C5a receptor (C5aR) or the empty plasmid vector as detailed previously [29]. U937 cells were cultured at 37°C and 5% CO₂ suspended in RPMI 1640 containing 10% FBS (HyClone, Logan, UT) and 400 µg/ml active G418 (Invitrogen Life Technologies, Carlsbad, CA), and maintained at a density between 2 × 10⁵ and 1.5 × 10⁶ per milliliter.

2.3 Construction of recombinant expressed DBP (reDBP) domains and DBP peptides

To construct GST fusion proteins, DNA sequences corresponding to the indicated DBP domains were amplified by PCR, designed to generate products with 5′ BamHI and 3′ XhoI restriction site. The E. coli expressed proteins were designated as follows: DBP domain I (residues 1–191), truncated (Δ) DBP domain I (ΔDI 1–125), truncated DBP domain I (ΔDI 1–175), DBP domain II (192–378), DBP domain III (379–458) DBP domains II & III (192–458), DBP domain II & ΔDIII (192–402), DBP domain II & ΔDIII (192–430). These were cloned into the corresponding site using pGEX-4T-2 expression plasmid as GST fusion partners in E. coli (sequences of all primers available from the authors upon request). E. coli strain XL-1 Blue (Stratagene, CA) served as the host for DNA manipulation and the E. coli strain BL-21 served the host for protein expression. Each construct was confirmed by DNA sequencing. DBP-derived peptides were synthesized and purified (>95% purity) by the American Peptide Company, Inc. (Sunnyvale, CA). Since all 28 cysteine residues in DBP form paired disulfide bonds, peptides were constructed to avoid free cysteine residues that could result in artifactual binding to cells. Peptides were designated as follows: Domain I peptide 1 (DIp1) residues 130–152 (EAFRKDPKEYANQFMWEYSTNYG); Domain I peptide 2 (DIp2) residues 150–172 (NYGQAPLSLLVSYTKSYLSMVGS); Domain III peptide 1 (DIIIp1) residues 378–390 (KELSSFIDKGQEL); Domain III peptide 2 (DIIIp2) residues 392–412 (ADYSENTFTEYKKKLAERLKA); Domain III peptide 3 (DIIIp3) residues 402–416 (YKKKLAERLKAKLPD); Domain III peptide 4 (DIIIp4) residues 414–434 (LPDATPKELAKLVNKRSDFAS); Domain III peptide 5 (DIIIp5) residues 438–445 (SINSPPLY); Domain III peptide 6 (DIIIp6) residues 447–458 (DSEIDAELKNIL). A complete list of truncated DBP domains and DBP-derived peptides and their properties is listed in Table I, the theoretical pI and molecular mass for each sequence was calculated using the Compute pI/MW tool available on the ExPASy website (http://ca.expasy.org/tools/pi_tool.html).

TABLE I.

Properties of Truncated (Δ) DBP and DBP Peptides Utilized in SPR Binding

Protein or Peptide	Symbol	Residue #	pI	Mr	Binds to CM5 Sensor Chip	Supports Cell Binding	Blocks Cell Binding to Native DBP
native DBP	DBP	1–458	5.22	51243	+	+	+
Domain I	DI	1–191	5.37	21454	+	+	+
Domain II	DII	192–378	4.88	20817	+	−	−
Domain III	DIII	379–458	4.89	9003	+	+	+
Δ Domain I (1–175)	DI (1–175)	1–175	5.17	19654	+	+	+
Δ Domain I (1–125)	DI (1–125)	1–125	5.25	13901	+	−	−
Domain II + DIII	DII + DIII	192–458	5.00	29778	+	+	+
Domain II + Δ DIII	DII + DIII (379–402)	192–402	4.75	23590	+	+	+
Domain II + Δ DIII	DII + DIII (379–430)	192–430	5.31	26720	+	+	+
Domain I peptide 1	DIp1	130–152	4.66	2875	+	−	−
Domain I peptide 2	DIp2	150–172	8.77	2479	+	+	+
Domain III peptide 1	DIIIp1	378–390	4.46	1494	−	N.D.	− alone + DIIIp2
Domain III peptide 2	DIIIp2	392–412	8.50	2505	−	N.D.	− alone + DIIIp1
Domain III peptide 3	DIIIp3	402–416	9.85	1800	−	N.D.	−
Domain III peptide 4	DIIIp4	414–434	8.63	2300	−	N.D.	−
Domain III peptide 5	DIIIp5	438–446	5.50	906	−	N.D.	−
Domain III peptide 6	DIIIp6	447–458	3.66	1463	−	N.D.	−

2.4 Expression and purification of reDBP domains

DBP domains were expressed in E. coli according to the protocol described by Swamy et al. [32]. BL21 cells carrying pGEX-4T-2 plasmids expressed fusion protein GST linked to the DBP constructs. E. coli were grown at room temperature (22–24°C) in Luria Broth (LB) containing 100 µg/ml ampicillin and 50 µg/ml chloramphenicol until absorbance at 600 nm (Abs 600) was 0.4–0.6. The expression of fusion proteins was induced by the addition of 0.5–1.0 mM IPTG for 4–8 hours. Cells were collected by centrifugation at 5000 × g. The E. coli pellets (from 1 liter) were resuspended in 40 ml TBS (50 mM Tris-HCl, pH 8.3, 150 mM NaCl) containing 0.2 mg/ml lysozyme, 0.1% Triton X-100 and 2 ml of a protease inhibitor cocktail mixture. The cells were disrupted by sonication and insoluble material was removed by centrifugation at 12,000 × g for 20 minutes. The detergent soluble cell lysate containing the expressed fusion protein was mixed with 1 ml of a 50% slurry of glutathione-agarose, preequilibrated in TBS. After several washes with TBS, the bound GST-DBP was eluted with 20 mM oxidized glutathione in 100 mM Tris Buffer, pH 8.3. The fusion protein subsequently was cleaved with thrombin and dialyzed against PBS to remove oxidized glutathione and GST was separated from reDBP using glutathione-sepharose affinity column. Sequencing fidelities of the reDBPs were confirmed by N-terminal sequencing.

2.5 Surface plasmon resonance (SPR) measurement of DBP-cell binding

The interactions between cells and DBP were evaluated using a BIAcore 2000 (BIAcore AB, Upsala, Sweden). Full-length native DBP, individual domains or DBP peptides were covalently coupled to a CM5 sensor chip using N-ethyl-N-(dimethylaminopropyl) carbodiimide / N-hydroxysuccinimide (EDC/NHS) according to the manufacturer’s instructions. The surface of the CM5 sensor chip was activated with EDC/NHS for 20 min before adding either DBP (5 µM) or DBP peptides (100 µM) in 10 mM sodium carbonate buffer, pH 5.0. Excess NHS was deactivated for 20 min using 1 M ethanolamine, pH 8.5. The efficiency of DBP coupling was determined by injecting 5 µg/ml affinity-purified goat anti-human DBP into the flow cell at 10 µl/min at 22°C. Cell-DBP binding interactions were determined by injecting cell suspensions at a flow rate of 5 µl/min at 22°C in Hanks’ balanced salt solution, pH 7.4 (HBSS) containing 0.005% Tween 20. The sensor chip was stripped and regenerated using 0.8 M glycine (pH 2.0) containing 0.6 M NaCl. The regeneration conditions were adjusted to achieve a subsequent binding response that was within 10% of the initial (first injection) binding value. A blank sensor chip that was EDC/NHS-activated and ethanolamine blocked was used as a background reference in all experiments. Net resonance response units (RU) were determined by subtraction of the background values using BIAevaluation software version 4.1. Each figure in the Results section shows a representative sensorgram of response units versus time, however, all experiments were repeated (minimum n ≥ 3) to verify results.

3. RESULTS

3.1 Measurement of cell binding to DBP by SPR

The experimental approach to determine the cell binding region(s) in DBP essentially was the same as previously utilized to identify the C5a chemotactic cofactor sequence [10]. Truncated forms of DBP were expressed in E. coli as GST fusion proteins, purified, and then utilized to determine if a particular truncated form could block cell binding to full-length native DBP in Biacore SPR. Overlapping peptides were synthesized once a region had been narrowed to about 50 amino acids. As a point of reference, figure 1 shows a schematic representation of the domain structure of DBP with the known functional regions highlighted. Figure 2A demonstrates that native DBP was successfully coupled to a Biacore CM5 sensor chip as detected by the molecular interaction between immobilized DBP and soluble anti-DBP. Figure 2B shows the dose dependent interaction between U937 cells and immobilized DBP. Cell concentrations below 10⁵/ml showed little increase in response units (RU), while cell concentrations higher than 2 × 10⁶/ml had an elevated basal RU after dissociation, making the sensor chip unsuitable for repeated usage. The cells used in this study were U937 cells stably transfected with the C5a receptor (U937-C5aR cells). Nevertheless, the kinetics of DBP binding between U937-C5aR, U937 cells transfected with an empty plasmid vector or wild-type U937 cells were identical (data not shown), further supporting our previous observation that DBP does not interact with the cell surface C5a receptor [33]. For optimal results all subsequent experiments used 10⁶ cells/ml interacting for 3 to 5 minutes at a flow rate of 10 µl/min. The increase in RU when cells are injected into the flow chamber was not due to nonspecific protein interactions because U937 cells did not interact with control chips containing either immobilized human β2-microglobulin or BSA (data not shown).

Figure 1.

Schematic representation of the DBP domain structure. The 458 amino acid sequence of human DBP with the three structural domains and known functional regions indicated. Domain I: amino acids 1–191; domain II: 192–378; domain III: 379–458; vitamin D binding: 35–49; C5a chemotactic cofactor: 130–149; G-actin binding: 373–403. The domain and functional regions are drawn approximately to scale.

Figure 2.

Initial characterization of DBP-cell binding by SPR. Purified human DBP was covalently coupled to a CM5 sensor chip. Panel A: detection of immobilized DBP with soluble anti-DBP. Panel B: overlay plot of sensorgrams when different U937 cell concentrations were injected into the flow cell.

3.2 Identification of cell binding regions in DBP Domains I and III using truncated proteins

The capacity of soluble DBP to block subsequent cell interactions with immobilized DBP was determined next. Figure 3 demonstrates that soluble native DBP added to U937 cells, prior to injection into the Biacore flow chamber, almost completely eliminated cell binding to immobilized DBP. Conversely, cells pretreated with purified human serum albumin were indistinguishable from control (buffer-treated) cells (data not shown), confirming the specificity of the DBP-cell interaction by SPR as we have previously reported using radioiodinated ligands [34]. Since full-length native DBP could effectively block cell binding to immobilized DBP, the effect of pretreating cells with individual recombinant expressed DBP domains was determined next. Figure 3 shows that pretreating cells with purified domain II did not alter subsequent cell interaction with DBP whereas domains I and III could block binding by approximately 50 and 70% respectively. To investigate if U937 cells can bind directly to purified DBP domains, each domain was individually coupled to a separate CM5 sensor chip. Figure 4A demonstrates that each domain was successfully coupled to the sensor chip as detected by the increase in RU with soluble anti-DBP. Figure 4B confirms that U937 cells do not bind to domain II but interact with both immobilized domain III and domain I. A truncated version of domain I containing amino acids 1–125 (as compared to full-length domain I at 1–191) also was not able to support cell binding whereas a 1–175 truncation showed binding identical to full-length domain I (Fig. 4B). These results indicate that the cell binding regions in DBP are located in the C-terminal part of domain I (126–175) and in domain III (379–458).

Figure 3.

Effect of soluble DBP or DBP domains on cell binding to immobilized DBP. U937 cells (10⁶/ml) in HBSS were pretreated with either buffer (control) or 0.2 µM of native DBP or the indicated DBP domain for 15 min at 22°C. Data is presented as an overlay plot of response curves obtained from the original sensorgrams.

Figure 4.

Cell binding to immobilized DBP domains. Purified recombinant DBP domains expressed in E. coli were individually coupled to separate CM5 sensor chips as described in Experimental Procedures. Panel A: detection of immobilized DBP domains with soluble anti-DBP. Panel B: overlay plot of response curves obtained from the original sensorgrams examining U937 cell binding to the indicated DBP domain.

3.3 Identification of a peptide cell localization sequence in DBP Domain I

The amino acid sequences responsible for mediating cell binding in domain I were investigated next using synthetic peptides. U937 cells were pretreated with either domain I peptide 1 (DIp1), amino acids 130–152, or domain I peptide 2 (DIp2), amino acids 150–172, and the binding to native DBP was subsequently evaluated. Figure 5A shows that DIp2 (150–172) could block cell binding to DBP by about 50% whereas DIp1 (130–152) was unable to prevent cells from interacting with immobilized DBP. Direct cell binding to immobilized peptides was determined after coupling each peptide to a separate sensor chip. Coupling of the peptides to the CM5 sensor chips was verified using anti-DBP (data not shown). Figure 5B confirms that only the DIp2 sequence (150–172) can support cell binding. Further analysis to identify the minimal amino acid sequence that supports cells binding, using three overlapping peptides that span the DIp2 sequence, was not successful (data not shown). These results indicate that the cell binding region in domain I is located in the peptide 2 (DIp2) sequence residues 150–172 (NYGQAPLSLLVSYTKSYLSMVGS).

Figure 5.

Cell binding capacity of domain I DBP peptides. Panel A: U937 cells (10⁶/ml) in HBSS were pretreated with either buffer (control), 0.2 µM of native DBP, 0.2 µM DIp1 (amino acids 130–152) or 0.2 µM DIp2 (amino acids 150–172) for 15 min at 22°C. Cells were then injected into the flow cell to interact with immobilized DBP. Data is presented as an overlay plot of response curves obtained from the original sensorgrams. Panel B: Purified domain I peptides (DIp1 and DIp2) were individually coupled to separate CM5 sensor chips as described in Experimental Procedures. Data is shown as an overlay plot of sensorgrams examining U937 cell binding to the indicated domain I DBP peptides.

3.4 Domain III cell binding sequence overlaps with the actin binding region in DBP

The cell binding region in domain III was investigated using a series of six peptides that spanned the entire 80 amino acids of the domain (Table I). However, we were unable to link these peptides to the CM5 sensor chip probably due to their small size and/or cationic pI’s (Table I). Therefore, we employed an alternative strategy utilizing recombinant expressed DBP with truncated versions of domain III attached to domain II. Domain II has superior immobilization to a CM5 sensor chip as detected by anti-DBP (Fig. 4A) but this domain does not support cell binding (Fig. 3 and Fig. 4B). Therefore, domain II serves as an excellent scaffold to evaluate cell binding in truncated versions of domain III. Four recombinant expressed DBPs were generated as GST fusion proteins in E. coli, domain II alone (192–378), full-length domains II and III (192–458), domain II plus truncated domain III (379–430), and domain II plus a greater truncation of domain III (379–402). All four proteins were individually immobilized to a CM5 sensor chip as evidenced by a robust increase in response units when the flow cell was injected with anti-DBP (data not shown). As noted previously (Fig. 3 and Fig. 4B) domain II alone did not support cell binding (Fig. 6A). However, Figure 6A demonstrates that all three forms of domain III support cell binding to the same extent, indicating that the binding sequence is located in the N-terminal part of the domain (379–402). Finally, this portion of domain III was analyzed further using the two peptides that span this sequence: DIIIp1 (378–390) and DIIIp2 (392–412). Figure 6B shows that when cells are pretreated with either peptide alone there is minimal (<10%) alteration of cell binding to immobilized domain II plus truncated domain III (379–402). In contrast, a combination of both peptides reduced subsequent cell binding by approximately 50% (Fig. 6B), indicating that the domain III cell binding region is localized to a sequence that spans DIIIp1 and DIIIp2. This sequence (379–402) largely overlaps with the actin binding sequence (373–403).

Figure 6.

Cell binding capacity of DBP domain III sequences. Panel A: overlay plot of response curves obtained from the original sensorgrams examining U937 cell binding to the indicated DBP domain. Panel B: U937 cells (10⁶/ml) in HBSS were pretreated with either buffer (control), 0.2 µM DIIIp1 (amino acids 378–390), 0.2 µM DIIIp2 (amino acids 392–412) or 0.2 µM of both peptides (DIIIp1+2) for 15 min at 22°C. Cells were then injected into the flow cell to interact with immobilized DII + ΔDIII (192–402). Data is presented as an overlay plot of response curves obtained from the original sensorgrams.

4. DISCUSSION

This paper provides the first evidence that human DBP contains distinct cell targeting sequences that may be required for its multiple functions. Results reveal two distinct cell binding sequences, one in the N-terminal domain adjacent to the C5a chemotactic cofactor region, and the other in the C-terminal domain overlapping with the G-actin binding sequence (Figure 7). Furthermore, analysis of the three-dimensional structure of DBP shows that the location of these sequences perhaps could allow DBP to function as an adaptor protein and bridge two distinct cell surface molecules. We propose that the cell binding sequences be named cb1 (for cell binding #1) for the N-terminal sequence and cb2 for the C-terminal sequence.

Figure 7.

Schematic representation of the DBP domain structure with proposed cell binding regions. The 458 amino acid sequence of human DBP with the three structural domains and known functional regions indicated. Domain I: amino acids 1–191; domain II: 192–378; domain III: 379–458; vitamin D binding: 35–49; C5a chemotactic cofactor: 130–149; cell binding #1 (cb1): 150–172; G-actin binding: 373–403; cell binding #2 (cb2): 379–402. The domain and functional regions are drawn approximately to scale.

DBP possesses at least four distinct and apparently unrelated functions: transport protein for vitamin D sterols, extracellular scavenger for G-actin released at sites of tissue injury, a cofactor that significantly enhances the chemotactic activity of C5a, and a macrophage and osteoclast activating factor [1, 2]. All of these functions require that DBP interact with the plasma membrane of its target cell. Indeed, numerous reports have demonstrated that DBP is associated with the plasma membrane of many different cell types [11–22], however, very little is known about this process. DBP interactions with the plasma membrane probably evolved in order to deliver vitamin D sterols to cells, whereas the other functions may be cell-specific adaptations using variations of the basic vitamin D delivery mechanism. Recently, our lab has begun to define the cell surface interactions required for DBP to function as a chemotactic cofactor for C5a. These reports demonstrate that several molecules are required for chemotaxis to complement-activated serum (DBP-dependent) but not to purified C5a (DBP-independent) [27, 30, 35] and suggest that a sequential multi-ligand assembly of a binding site may be necessary for DBP to mediate this function. However, detecting cell surface DBP ligands, that may interact only transiently and with low avidity, is an exceedingly difficult challenge. Thus, we investigated the converse approach of identifying cell binding regions in DBP utilizing Biacore SPR to measure the relative binding of U937 cells to immobilized DBP. This technique is superior to employing DBP covalently coupled to either ¹²⁵I [24] or Alexafluor-488 [27] because it evaluates binding in real time. In addition, controls and experimental samples can be analyzed simultaneously because of the four channel parallel arrangement of Biacore sensor chips. Consequently, this assay is considerably more sensitive and the results more reproducible than previously used techniques [24, 27].

The location of the cell binding regions in DBP may hint at their functional importance. The N-terminal cell binding region (residues 150–172) is immediately adjacent to the C5a chemotactic cofactor sequence (residues 130–149) and may be needed to localize this sequence to its cell surface target (Fig. 7). Previously, we have shown that the C5a chemotactic cofactor sequence peptide can effectively block full-length DBP from enhancing chemotaxis to C5a [10]. However, the peptide by itself is unable to augment chemotaxis to C5a perhaps because it requires the adjacent cell surface localization sequence (residues 150–172). The cell binding sequences may be required for cell delivery of vitamin D sterols. Saturation of DBP with vitamin D (either the 25-hydroxy or 1,25-dihydroxy forms) does not appear to interfere with binding to neutrophils or the myeloid cell lines HL-60 and U937 [30]. Active vitamin D (1,25 dihydroxy vitamin D) mediates cellular effects via the vitamin D receptor (VDR), a well characterized nuclear receptor [36], but also has been shown to trigger rapid signaling events after binding to the plasma membrane [36]. This rapid cell surface effect of vitamin D is thought to be mediated via a protein known as 1,25-D membrane associated rapid response steroid (MARRS) receptor, also know as ERp57, GRp58 or ERp60 [37, 38]. It is interesting to speculate that DBP also may interact with the cell surface MARRS receptor for delivery of vitamin D. Indeed, we have reported that active vitamin D (but not the inactive 25-hydroxy form), when bound to DBP, completely eliminates the C5a chemotactic cofactor function, an effect that is dependent upon cell surface alkaline phosphatase (AP) activity [30]. Furthermore, studies have shown that both cell surface AP [39] and perhaps the MARRS receptor [40] interact with annexin A2, a molecule that we previously demonstrated binds DBP and is required for C5a chemotactic cofactor activity [27]. Thus, it is conceivable that cell localization sequences in DBP target the protein to a cell surface protein complex for delivery of vitamin D sterols.

The other major binding function of DBP is to bind and remove G-actin from extracellular fluids. DBP-actin complexes have been detected in the blood of individuals with a variety of conditions including pregnancy, adult respiratory distress syndrome, sepsis, bacterial pneumonia, hepatitis, acetaminophen overdose, and multiple trauma cases [1, 2]. These complexes are rapidly cleared from the circulation primarily by the liver [1, 2]. Several animal studies (using rats and rabbits) have demonstrated that DBP-actin complexes are removed from the blood 5–6 times faster than DBP alone [41–43]. Thus, DBP-actin complexes may be thought of as a damage-associated molecular pattern (DAMP), similar to other intracellular molecules released from necrotic cells such as HMGB-1, and the combination of cell binding sequences in DBP and actin target the complex to hepatic scavenger receptors for removal. It is interesting to note that the cell binding region in the C-terminal domain (residues 379–402) overlaps with much of the actin binding site (residues 373–403). However, a previous study from our lab has shown that is no difference in cell (neutrophil) binding between DBP alone and DBP-actin complexes, and actin bound to DBP does not alter the C5a chemotactic cofactor function [30]. The deduced linear sequences of both the actin and cell binding regions are rather large, but the precise amino acid sequences required for these functions actually may be considerably smaller and possibly do not overlap. Another possibility is that DBP may bind actin on the cell surface. The initial binding of DBP to cells is relatively weak but once it has reached a steady-state plateau between binding and shedding (30 min at 37°C) DBP is bound tightly to the detergent insoluble fraction of the plasma membrane and can only be dissociated by harsh conditions (1% SDS or 0.1 M carbonate, pH 11) [24, 34]. This would suggest that DBP binds tightly to at least one of its ligands, and DBP is known to bind G-actin with high affinity (10⁻⁹ M). Actin is the most prevalent intracellular protein, however, several investigators have reported actin on the extracellular face of the plasma membrane in lymphocytes and endothelial cells [11, 13, 44–46]. Thus, it is conceivable that the tight DBP-membrane binding is due to the interaction of DBP with cell surface actin. Indeed, we have observed DBP bound to actin on the plasma membrane of both neutrophils and the myeloid cell line U937 (L.A. McVoy and R.R. Kew, manuscript in preparation). The reason that actin is expressed on the extracellular face of the plasma membrane is not clear but may reflect either cell activation or a pre-apoptotic state.

The significance of DBP-cell surface interactions in the functions of this protein has not been appreciated fully. This study may help to facilitate future investigations designed to address these questions. The results reported herein clearly demonstrate that there are only two cell binding regions in DBP. However, the major caveat with this study is that the two cell localization sequences have been defined for leukocytes of myeloid origin and DBP may possess additional binding regions for other cell types. For example, megalin and cubulin have been shown to bind DBP [25, 26], but these multi-ligand receptors are expressed largely on epithelial cells, therefore, it is not know if these cell types utilize the same cb1 and cb2 sequences to bind DBP. Currently, our lab is expressing recombinant DBP with deletions in the cb1, cb2 or both sequences and will use these proteins to screen the cell binding capacity of a variety of cell types. In summary, human DBP contains two cell surface localization sequences that may permit the protein to function as an adapter molecule and bridge two distinct molecules.

Abbrevations used

DBP

vitamin D binding protein

DAMP

damage-associated molecular pattern

MARRS

membrane associated rapid response steroid receptor

RU

response units

SPR

surface plasmon resonance

U937-C5aR

U937 cell line transfected with the C5a receptor