A negatively charged cluster in the disordered acidic domain of GPIHBP1 provides selectivity in the interaction with lipoprotein lipase

doi:10.21203/rs.3.rs-4348425/v1

Download PDF

Article

A negatively charged cluster in the disordered acidic domain of GPIHBP1 provides selectivity in the interaction with lipoprotein lipase

https://doi.org/10.21203/rs.3.rs-4348425/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 23 Aug, 2024

Read the published version in Scientific Reports →

You are reading this latest preprint version

GPIHBP1 is a membrane protein of endothelial cells that transports lipoprotein lipase (LPL), the key enzyme in plasma triglyceride metabolism, from the interstitial space to its site of action on the capillary lumen. An intrinsically disordered highly negatively charged N-terminal domain of GPIHBP1 contributes to the interaction with LPL. In this work, we investigated whether the plethora of heparin-binding proteins with positively charged regions found in human plasma affect this interaction. We also wanted to know whether the role of the N-terminal domain is purely non-specific and supportive for the interaction between LPL and full-length GPIHBP1, or whether it participates in the specific recognition mechanism. Using surface plasmon resonance, affinity chromatography, and FRET, we were unable to identify any plasma component, besides LPL, that bound the N-terminus with detectable affinity or affected its interaction with LPL. By examining different synthetic peptides, we show that the high affinity of the LPL/N-terminal domain interaction is ensured by at least ten negatively charged residues, among which at least six must sequentially arranged. We conclude that the association of LPL with the N-terminal domain of GPIHBP1 is highly specific and human plasma does not contain components that significantly affect this complex.

Biological sciences/Biochemistry/Enzymes/Hydrolases

Biological sciences/Biochemistry/Lipids/Lipoproteins

Biological sciences/Biochemistry/Enzyme mechanisms

GPIHBP1 is a capillary endothelial cell protein that transports lipoprotein lipase (LPL), the main enzyme in hydrolytic degradation of blood triglycerides, across endothelial cells to the capillary lumen. Biallelic loss-of-function mutations in GPIHBP1 cause LPL to be retained in the interstitial space, resulting in severe hypertriglyceridemia, which in turn can lead to life-threatening acute pancreatitis[1]. According to a recent study by Song et al.[2], part of the GPIHBP1 transported LPL relocates to heparan sulfate proteoglycans (HSPGs) in the glycocalyx. At this site, LPL acts on triglyceride-rich lipoproteins, breaking down their triglycerides into fatty acids and monoglycerides that are usable components for cells. It is not known whether the displacement of LPL from GPIHBP1 to the glycocalyx is entirely spontaneous or is influenced by some plasma component. Understanding this mechanism may be important because its misregulation can lead to elevations in plasma triglyceride levels, which are correlated with an increased risk of cardiovascular disease and other metabolic disorders.

Two distinct regions of GPIHBP1 - an intrinsically disordered N-terminal acidic domain with a long stretch of aspartate and glutamate residues, and a cysteine-rich Ly6 domain with a three-finger structure – are involved in the interaction with LPL. Both GPIHBP1 domains are also required for its transport function[3]. Surface plasmon resonance (SPR) studies show that while the interaction of the N-terminal domain with LPL is characterized by a high on-off rate, the complex with Ly6 forms and also dissociates at a slower rate[4–6]. While the sequence of the Ly6 domain is highly conserved between species, the N-terminal domains differ in length, distribution, and number of negative charges[7]. For example, human N-terminal domain of GPIHBP1 contains 21 negatively charged residues, but mouse or opossum have 17 or 32 acidic residues, respectively. X-ray structures of the LPL/GPIHBP1 complex reveal that the Ly6 domain binds to the C-terminal domain of LPL and that this association is mainly hydrophobic in nature[8, 9]. Although the binding region of LPL involved in the interaction with the N-terminal domain was not identified in the X-ray structures, it is probable that the highly positively charged region of LPL plays a crucial role in its formation. This region is also likely to be involved in LPL’s interaction with other polyanions such as the polysaccharides heparin and heparan sulfates[10]. Comparison of affinities of synthetic peptides corresponding to the sequence of the N-terminal domain of mouse, bovine or human GPIHBP1 suggested that the number of negatively charged residues in the N-terminal domain might be important[4]. The GPIHBP1 N-terminal domain of several species also contains a tyrosine whose sulfation may play a role in affinity for LPL[6]. However, the structural motifs of the N-terminal domain of GPIHBP1 that confer its high affinity to LPL remain largely unclear.

Previous studies have shown that LPL binds strongly to heparin and heparan sulfates and, like the N-terminal domain, this interaction is transient and characterized by high on- and off-rates[11]. Furthermore, the interaction seems to not be highly specific, LPL prefers HS sequences with high negative charge density[12, 13]. These observations raise a question of specificity for the binding of the N-terminal domain to LPL. Is it completely non-specific, having only an additional role in the interaction with full-length GPIHBP1 or does it play a specific role? It is noteworthy to bear in mind that many heparin-binding proteins have been identified in human plasma[14–16], the common feature of which is the presence of a positively charged region in their folded structure[17]. Considering that the interaction of some proteins with heparin is primarily non-specific and ionic, depending mainly on the charge density, it is reasonable to ask whether the N-terminal domain of GPIHBP1 could be a potential ligand for them.

Here, we investigated whether human plasma contains components that bind to the N-terminal domain of GPIHBP1 or affect its interaction with LPL. Using SPR, affinity chromatography, and fluorescence techniques, we were unable to identify any plasma component, other than LPL, that bound the N-terminus with detectable affinity or affected its interaction with LPL. In contrast, we ascertain several plasma proteins that bound avidly to heparan sulfate or heparin. Based on these observations, we conclude that the association of LPL with the N-terminal domain of GPIHBP1 is highly specific and human plasma does not contain components that substantially affect this complex. By combining binding and stabilization studies of different synthetic peptides, we show that the high affinity of the LPL/N-terminal domain interaction is ensured by at least ten negatively charged residues, among which at least six must be in a sequentially arranged cluster.

Reagents

LPL was purified from bovine milk as previously described[18]. Aliquoted stock solutions of LPL were stored at -80°C in a buffer containing 10 mM Bis-Tris, pH 6.5, 1.5 M NaCl. Samples were used immediately after thawing and only used once. Synthetic peptides (Table 1) were purchased from GeneCust (Luxembourg). Some peptides contained an extra cysteine residue at the C-terminal end. The synthetic peptides were biotinylated and labeled by DyLight 488 dye (Pierce) at cysteines as previously described[4]. Heparan sulfate was biotinylated at amino groups as previously described[11]. Extinction coefficients at 280 nm for determination of peptide/protein concentrations were as follows: LPL – 70440 M^− 1 cm^− 1; human N-terminal peptide, bovine N-terminal peptide, peptide 1 and peptide 4–1480 M^− 1 cm^− 1; mouse N-terminal peptide – 2960 M^− 1 cm^− 1. The extinction coefficients were calculated according to Gill and Hippel[19]. Concentrations of LPL were calculated using its monomer molecular weight of 55 kDa. 1,2-Di-O-lauryl-rac-glycero3-glutaric acid 6-methylresorufin ester (DGGR) (#30058) was purchased from Sigma Aldrich. Nonfasting human EDTA-plasmas were purchased from the Tallinn Blood Centrum. Lipoprotein deficient human plasma was obtained from plasma by using an ultracentrifuge as previously described[20]. Triglyceride and cholesterol concentrations were determined using Triglyceride Colorimetric Assay Kit (Cayman, USA) and Cholesterol Fluorometric Assay Kit (Cayman, USA).

Surface plasmon resonance experiments.

SPR experiments were conducted on a Biacore 3000 (GE Healthcare Life Sciences) instrument. Neutravidin (Sigma) was covalently attached to the surface of CM5 (GE Healthcare Life Sciences) sensor chips using the amine coupling kit (GE Healthcare Life Sciences). To study the binding of human plasma components to different ligands, biotinylated N-terminal peptide of GPIHBP1 or biotinylated heparan sulfate were bound to the sensor chip surface via biotin-NeutrAvidin interaction. 578 RU of biotinylated N-terminal GPIHBP1 was bound to the surface of the sensor which corresponded to a surface density of 148.5 fmol/mm².

Nonfasting pooled human plasma samples with a mean TG concentration of 1.08 mM (n = 8) or pooled lipoprotein deficient human plasma samples (n = 3) were diluted 10-fold and injected over the surfaces (60 µl, 20 µl/min). In control experiments, 20 nM, 40 nM or 80 nM purified bovine LPL was added to the same 10-fold diluted pooled lipoprotein deficient human plasma. Measurements were performed at 25°C in a running buffer containing 20 mM HEPES, pH 7.4, 0.15 M NaCl.

The affinity between various peptides and LPL was assessed using a SPR competition assay, following the experiments first described by Reimund et al[4]. The sequences of the investigated peptides are presented in Table 1. The experimental steps of the binding study were as follows: 1) Biotinylated N-terminal peptide of GPIHBP1 was immobilized to the sensor chip via NeutrAvidin. 2) LPL (800 nM) was mixed with different peptides at increasing concentrations at 4°C. LPL or solutions of LPL/peptide complexes were injected over the sensor chip’s surface (30 µl, 5 µl/min). 3) Binding of LPL to the immobilized N-terminal peptide of GPIHBP1 was registered near the equilibrium at the end of each injection. Measurements were carried out at 4°C. Running buffer contained 20 mM NaH₂PO₄, pH 7.4, 2 mg/ml BSA, 0.4 M NaCl. The affinity of the peptides for LPL, expressed as K_D values, was calculated using Eq. (1).

$$\varDelta R=\frac{a\bullet {K}_{D}}{{K}_{D}+{P}_{0}} \left(1\right)$$

where P₀ is the concentration of peptide in the injected solution, the K_D is the equilibrium dissociation constant, ΔR is the change in SPR at equilibrium, and ‘a’ is the change in SPR for binding of LPL to the surface when the solution did not contain free peptide.

Direct binding studies were performed to investigate the interaction between peptide 6 and LPL (Table 1). Biotinylated peptide 6 was attached to the surface via NeutrAvidin and solutions of 1, 2 or 3 µM LPL were injected over this surface. Experimental conditions were the same as in competition experiments.

Table 1

Peptides of the N-terminal region of GPIHBP1
Peptide	Sequence	Negatively charged amino acid residues
Human N-terminal peptide	QQEEEEEDEDHGPDDYDEEDEDEVEEEETC	21
Peptide 1	QQEEEEEDEDHGPDDY	10
Peptide 2	EDHGPDDYDEEDED	10
Peptide 3	EEDEDEVEEEETC	10
Peptide 4	EVEEEETC	5
Peptide 5	HGPDDYDEEDEDEVEEEETC	13
Peptide 6	QQEAEAAEEAHDGPEDYAEADADEVAEAETC	13
Peptide 7	VDQDYEDDHETEDDEEQDPEEEEGEEEEEC	21
Peptide 8	EDEVEEEETC	7
Mouse N-terminal peptide	AQEDGDADPEPENYNYDDDDDEEEEEETC	17
Bovine N-terminal peptide	AQEDEDDDPDAGREGYDDEDEEEEEA	17

Affinity chromatography.

Fasting human plasma (17 ml) was loaded to an 8 ml heparin column (Heparin Sepharose 6 Fast Flow affinity resin, GE Healthcare Life Sciences) in a buffer containing 100 mM TRIS, pH 7.4, 0.15 M NaCl. The column was washed with 80 ml of the same buffer. Heparin binding proteins (i.e. bound material) were eluted from the column using a buffer containing 100 mM TRIS, pH 7.4, 2 M NaCl. Eluted proteins were pooled together, dialyzed to buffer containing 100 mM TRIS, pH 7.4, 0.15 M NaCl and loaded to an affinity column containing immobilized N-terminal peptide of GPIHBP1. The latter was made by attaching biotinylated N-terminal peptide to HiTrap Streptavidin HP column (GE Healthcare Life Sciences). The column was washed with 100 mM TRIS, pH 7.4, 0.15 M NaCl buffer and elution was performed in a 0.15–2 M NaCl gradient. All steps were carried out at 10°C. Identification of human plasma proteins that bound to the heparin column was obtained as a service from the Proteomics Core Facility at the Institute of Technology, University of Tartu (Tartu, Estonia).

Fluorescence anisotropy.

Fluorescence anisotropy experiments were conducted on a Hitachi F-7000 (Hitachi High-Tech, Japan) fluorescence spectrophotometer. The excitation and emission wavelengths were 493 nm and 518 nm, respectively. Experiments were done either in phosphate buffer (20 mM phosphate, pH 7.4, 0.15 M NaCl), or in the same buffer with added 10 mg/ml BSA or 50% lipoprotein free human plasma. Dylight 488 maleimide labeled N-terminal peptide of GPIHBP1 at concentrations of 100 nM were mixed with increasing concentrations of LPL. Next, fluorescence was measured when excitation and emission polarizers both were oriented vertically, and when the excitation polarizer was oriented vertically, and the emission polarizer was oriented horizontally.

From the acquired data the fluorescence anisotropy was calculated using Eq. (2):

$$r=\frac{{I}_{\parallel }-{I}_{\perp }}{{I}_{\parallel }+{2I}_{\perp }} \left(2\right)$$

where r is fluorescence anisotropy, ${I}_{\parallel }$ is observed fluorescence intensity when the emission polarizer was oriented parallel to the direction of the polarized excitation, ${I}_{\perp }$ is the observed fluorescence intensity when the emission polarizer was oriented perpendicular to the direction of the polarized excitation.

Dependence between change of anisotropy and LPL concentration was fitted with Eq. (3) for calculation of the K_D values for a 1:1 binding model:

$$r={r}_{max}\left({L}_{0}+{P}_{0}+{K}_{D}-\sqrt{{\left({L}_{0}+{P}_{0}+{K}_{D}\right)}^{2}-4{L}_{0}{P}_{0}}\right) \left(3\right)$$

where r is the change of fluorescence anisotropy, r_max is the value of anisotropy of the complex of LPL with the N-terminal peptide, L₀ is the concentration of LPL, P₀ is the concentration of the peptide, K_D is the equilibrium dissociation constant.

Determination of LPL activity with DGGR

Fluorescence spectroscopy was utilized to assess the effects of the N-terminal GPIHBP1 and synthetic peptides on the thermostability of LPL by monitoring LPL activity using the fluorogenic substrate DGGR. Experiments were conducted using a spectrofluorophotometer (Shimadzu RF-5301 PC, Shimadzu Corporation, Japan) at an excitation wavelength of 572 nm and an emission wavelength of 605 nm. 10 nM LPL was incubated at 37°C in 20 mM HEPES, pH 7.4, 150 mM NaCl either alone, with 1 µM nGPIHBP1, or with 1 µM synthetic peptides. Initial LPL activity was measured at 37°C immediately after mixing, and then at 10-minute intervals.

Modeling of the predicted complex

The cryo-EM structure of LPL for the visualization of positively charged surfaces was obtained from the PDB entry 8ERL[21]. The structure of LPL in complex with either full-length GPIHBP1 or the N-terminal domain of GPIHBP1 was predicted using ColabFold[22], which uses an optimized approach to accelerate protein-protein complex modeling by Alphafold2-Multimer[23]. The human LPL and GPIHBP1 sequences were obtained from Uniprot database entries Q6IAV0 and Q8IV16, respectively[24]. Five predicted structures were generated for each complex based on the chosen sequences and the model with the highest confidence score was chosen for visualization. ChimeraX was used to create the figures and calculate the Coulombic electrostatic potential of predicted structure surfaces[25].

LPL is the only plasma component that binds the N-terminal domain of GPIHBP1 with high affinity

To investigate whether human plasma contains proteins in addition to LPL that bind to the N-terminal domain of GPIHBP1, we performed binding studies using SPR. Solutions of pooled human plasma samples (n = 8) or pooled lipoprotein-deficient human plasma samples (n = 3) were injected into sensor chip flow cells pre-immobilized with the N-terminal domain of GPIHBP1 (Fig. 1) or with heparan sulfate (Fig. 2). To assess the effect of nonspecific binding, the plasma solutions were injected into the flow cells that did not contain the N-terminal domain or heparan sulfate. The results of these experiments suggested that neither lipoprotein-deficient nor lipoprotein-containing human plasma contained components that bound to the N-terminal domain of GPIHBP1 with detectable affinity. In the case of lipoprotein-containing plasma, nonspecific binding to the sensor chip surface was even higher than to the N-terminal domain (Fig. 1b). However, when 20 nM, 40 nM or 80 nM LPL was present in the lipoprotein deficient plasma samples, association with the N-terminal domain of GPIHBP1 was evident, and increased when the concentration of LPL was raised (Fig. 1c, Fig. 1d and Fig. 1e).

In contrast to the N-terminus of GPIHBP1, considerable binding of human plasma components to the immobilized heparan sulfate was observed (Fig. 2a and Fig. 2b), consistent with previous studies showing that human plasma contains numerous heparin-binding proteins[14–16]. These experiments suggested that even though human plasma contains many heparin binding proteins, they either do not interact with the N-terminal domain of GPIHBP1, or their affinity to this domain is much lower than that of LPL.

In addition to SPR experiments, we used affinity chromatography to study the binding of plasma components to the N-terminal domain of GPIHPB1 and heparin. We used a streptavidin-Sepharose column onto which biotinylated N-terminal of GPIHBP1 was immobilized. Heparin was directly immobilized to CNBr activated agarose. Mass spectrometry identified a total of 76 plasma proteins, which were bound to heparin-agarose and subsequently eluted using a NaCl concentration gradient. The strongest proteolytic peptide intensity signals were given by antithrombin, kallikrein, thrombin, apolipoprotein E, histidine rich glycoprotein, alpha1 microglobulin, coagulation factor XI, complement factor D. However, these proteins did not bind to the N-terminal domain of the GPIHBP1 affinity column (Fig. 3a) — all these proteins were present in the flow-through and no additional proteins were eluted in the NaCl gradient. In contrast, LPL in the presence of BSA bound to the column with immobilized N-terminal GPIHPB1 and eluted at NaCl concentration between 0.4–0.6 M (Fig. 3b).

Components of human plasma do not substantially interfere with the interaction between LPL and the N-terminal domain of GPIHBP1

To further investigate the interaction between LPL and the N-terminal domain of GPIHBP1, we used fluorescence anisotropy which allowed to conduct studies in undiluted blood plasma, i.e. the environment in which LPL functions in vivo. For comparison, measurements were performed under three different conditions: 1) in a standard phosphate buffer, 2) in the same buffer with added 10 mg/ml BSA or 3) in the same buffer which contained lipoprotein free human plasma. As can be seen in Fig. 4, the affinity of the interaction was the highest in phosphate buffer (K_D = 60 nM). In buffers with BSA or lipoprotein free plasma the interaction was slightly weaker. However, in these cases a simple binding model did not fit the data well and the K_D value could not be reliably determined. This suggests a more complex binding mechanism. The comparable effect of BSA and plasma on the interaction suggests that albumin acted as an effector, whereas other components of human plasma did not appear to influence this interaction. The weak binding of BSA to LPL as shown in our recently published study would explain the slightly reduced affinity and the complex interaction mechanism[26]. These observations are in line with SPR findings in diluted human plasma (Fig. 1) showing the specificity and selectivity of this interaction and suggest that other plasma components do not compete with LPL to interact with the N-terminal domain of GPIHBP1.

A cluster of negatively charged residues drives specificity of the interaction between the N-terminal domain of GPIHBP1 and LPL

The observations that binding of LPL to the N-terminal domain of GPIHBP1 is specific prompted us to ask what structural features of the N-terminal domain ensure the high affinity of this interaction (Fig. 5). Firstly, we examined whether various regions of the domain may have different contributions to the total affinity. We chose peptides 1 and 2 (see Table 1), whose sequences partially overlapped but represented distinct regions within the human N-terminal domain of GPIHBP1. These peptides contained the same number (10) of negatively charged residues. Their affinities for LPL were compared to peptide 3 which also contained 10 negatively charged residues and whose affinity has been already previously determined[4]. Together the sequences of these three peptides make up the entire N-terminal domain sequence. All three peptides bound to LPL with affinities that did not differ greatly. However, peptide 3 had the highest affinity and peptide 1 had the lowest affinity. Their comparable affinities for LPL suggested that all parts of the N-terminal domain of GPIHBP1 contribute to the interaction with LPL.

Next, we examined how important is the number and distribution of negatively charged groups. We designed peptide 6, the sequence of which was obtained by replacing several aspartate and glutamate residues with alanine in the sequence of the human N-terminal domain of GPIHBP1. As a result of such substitutions, there were no regions in peptide 6 with more than two adjacent negatively charged residues. The 21 amino acid sequence of peptide 6 contained 13 negatively charged groups, as did peptide 5, whose negatively charged residues were arranged compactly and formed a cluster of seven negatively charged residues in the middle of its 15 amino acid sequence. Peptide 5 exhibited a binding affinity to LPL comparable to that of the full-length N-terminus. Conversely, peptide 6 failed to demonstrate detectable binding under the experimental conditions employed. These observations suggest either considerably reduced affinity to LPL or a complete lack of interaction. Thus, the presence of only a certain number of negatively charged residues alone does not guarantee high affinity binding to LPL. Instead, the sequential arrangement of the negatively charged group in the sequence emerges as the critical factor. At the same time, the presence of negatively charged clusters was also not sufficient, because short cluster-containing peptides 4 and 8 did not interact with LPL. Summarizing the obtained results on Fig. 5, our findings indicate that for robust binding to LPL, it is essential for a sequence to contain at least one negatively charged cluster of 6 residues, along with a minimum total of 10 negatively charged residues.

Clusters of negatively charged residues are crucial for LPL stabilization

As previous studies have shown, the N-terminal domain of GPIHBP1 stabilizes LPL from spontaneous thermal inactivation[5, 27]. We therefore investigated whether the same peptides that were used in the interactions study could also prevent loss of catalytic activity of LPL at 37°C (Fig. 6 and Supplementary Fig. S1). Peptide 4, which has the shortest sequence and no detectable affinity to LPL, could not stabilize LPL compared to plain buffer. Peptide 8, with two additional negatively charged residues compared to peptide 4, also failed to exert any effect on the stability of LPL. However, as the number of negatively charged residues increased further, so did the effect of the peptides on the stability of LPL. Peptide 3 conserved 60% of LPL activity. Peptide 5, and the human N-terminal peptide, which had a higher affinity to LPL compared to peptide 3, increased LPL stability even further. Interestingly, while peptide 6 has the same number of negatively charged residues as peptide 5, it failed to increase LPL stability compared to plain buffer. This demonstrates that, much like the binding affinity of peptides to LPL is not solely dependent on the number of negatively charged residues, neither is the effect of peptides on LPL stability.

Predicted models show interaction of the N-terminal domain of GPIHBP1 with the large positively charged region of LPL

LPL has a large positively charged region spanning from the C-terminal domain, across the hinge region, to the N-terminal domain as shown on Fig. 7a. This patch is comprised of four distinct positive charge clusters of which three are important for binding to heparin[10]. We hypothesized that the same clusters might be important for the interaction with the N-terminal domain of GPIHBP1. While the binding of the Ly6 domain of GPIHBP1 has been elucidated[8, 9], the binding of the N-terminal domain has remained elusive, likely due to its disordered and dynamic nature. We therefore used ColabFold to predict the localization of the interaction interface between the N-terminal domain of GPIHBP1 and LPL (Fig. 7b). All predicted models consistently demonstrated binding of the highly negatively charged N-terminal domain of GPIHBP1 to the basic patch of LPL. This was also true in the case of full-length GPIHBP1 (Fig. 7c), where the Ly6 domain was simultaneously bound to the same C-terminal region of LPL as shown in crystal structures[8, 9]. In both complexes shown in Fig. 7, the C-terminal segment of the N-terminal domain, which contains a cluster of negatively charged groups (DEEDEDEVEEEE), directly interacts with the surface of LPL. The rest of the domain remains away and can support interactions with long-range electrostatic forces. However, it should be considered that the per-residue model confidence score (pLDDT) for the predicted structure of the complexes was very low (< 50), which means that interactions between other regions are also possible. It is also feasible that the low scores are a consequence of the inherent disorder of the N-terminal GPIHBP1 peptide and are a predictor of the dynamic nature of the peptide, rather than a sign of low prediction confidence[28].

In the present study, we show that although plasma contains numerous proteins that bind to the negatively charged polysaccharides heparin and heparan sulfate, we were unable to identify any that interacted with the highly negatively charged and disordered N-terminal domain of GPIHBP1 with detectable affinity. At the same time LPL binds to this domain with high affinity[4, 6]. Furthermore, plasma proteins did not block the interaction between LPL and the N-terminal domain in 10-fold diluted plasma as shown by SPR. This observation supports the speculation that movement of LPL from GPIHBP1 to HSPGs in the glycocalyx might occur spontaneously[2]. Fluorescence anisotropy measurements in undiluted human plasma suggested that only albumin, the major plasma protein, slightly reduced this interaction. Such a large difference between LPL and other heparin-binding proteins was somewhat unexpected because several of the plasma proteins interact with heparin with considerable affinity[17]. Moreover, the interaction of several of these proteins with heparin has been shown to be rather non-specific without favoring specific heparin sequences. In addition, plasma concentrations of some heparin-binding plasma proteins are relatively high. A prime illustration of this concept can be found in thrombin, which typically maintains a plasma concentration ranging between 5–10 mg/dl or 0.7–1.3 mM[29] and its interaction with heparin is predominantly nonspecific[30]. Thus, it is difficult to find an explanation for the large difference between LPL and other heparin-binding proteins. We speculate that the high affinity of LPL/N-terminal domain interaction is due to the large positively charged polyanion binding region in LPL (∼2400 Å²)[8], which includes the region in the C-terminal domain, the hinge region, and the N-terminal catalytic domain of the enzyme. In other heparin-binding proteins, the area of the corresponding regions is substantially smaller, ranging from 200-1150 Å[17]. The wide (∼60x40Å²)[8] and positively charged region in LPL may allow the N-terminal domain to be positioned there in several different ways, leading to more degrees of freedom, an increase in binding entropy and thus a potential increase in the binding affinity[31, 32].

The results of the present study suggest that the high affinity of the N-terminal domain for LPL is ensured by the presence of at least 10 negatively charged residues, at least 6 of which are in a consecutive cluster. This conclusion is also supported by the knowledge that all known GPIHBP1 sequences from different species have both a cluster and a sufficient number of negative residues in their N-terminal domain. The important role of the cluster was particularly evident in the comparison of peptides 5 and 6 which had an equal number of negative charges but showed a considerable difference in their interaction with LPL. Only the cluster-containing peptide 5 bound to LPL with substantial affinity and stabilized its active conformation. According to the small-angle X-ray scattering (SAXS) analyses[6], the chain length of the disordered N-terminal domain is 68 Å, most of which could fit into the highly positively charged polyanion-binding region in LPL. However, our results suggest that the high density of negative charge of the cluster region is a prerequisite for strong binding. Therefore, the cluster region is likely to be the "hot" region of the N-terminal domain in the interaction with LPL. Our data further indicates a strong correlation between the affinity of the examined peptides and their capacity to stabilize LPL. It is reasonable to infer that the stabilization of LPL necessitates a dense concentration of negatively charged sequences within the polypeptide chain, facilitating binding to regions of LPL characterized by a high density of positive charge. Electrostatic repulsion forces likely underlie LPL's instability, arising from positively charged residues within LPL's highly positively charged region.

Our modeling findings align with the SPR experiments and stabilization studies: the cluster situated at the C-terminus of the N-terminal domain EEDEDEVEEEETC is projected to establish direct contact with the LPL region positioned at the junction of the C-terminal domain and hinge region. Other segments within the N-terminal domain are not directly engaged. Their influence on the interaction likely stems from long-range electrostatic effects. All examined peptides demonstrated interaction with this region in the modeled complexes.

In summary, in this report we show that the interaction between the strongly negatively charged disordered N-terminal domain of GPIHBP1 and the natively folded LPL is highly specific in human plasma environment. The sequential cluster of negatively charged residues ensures this specificity. Other plasma components do not substantially influence this interaction.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

ACKNOWLEDGEMENTS

This work was supported by TallinnUniversity of Technology (grant SS22005 to A.L).

AUTHOR CONTRIBUTIONS

A.L. and M.R. conceptualized the study. M.R. performed the experiments shown in Figs. 1, 2, 3, 4. R.R. compiled the data for Fig. 5. R.R and N.-N.S. performed the experiments shown in Fig. 6. R.R. performed the structural modeling on Fig. 7. A.L. acquired funding for the project and supervised the project. R.R, M.R, A.L. wrote the main manuscript text. All authors reviewed the manuscript.

COMPETING INTERESTS

The authors declare no competing interests.

Davies, B. S. J. et al. GPIHBP1 Is Responsible for the Entry of Lipoprotein Lipase into Capillaries. Cell metabolism 12, doi:10.1016/j.cmet.2010.04.016 (2010).
Song, W. et al. The lipoprotein lipase that is shuttled into capillaries by GPIHBP1 enters the glycocalyx where it mediates lipoprotein processing. Proceedings of the National Academy of Sciences 120, doi:10.1073/pnas.2313825120 (2023).
Song, W. et al. Electrostatic sheathing of lipoprotein lipase is essential for its movement across capillary endothelial cells. Journal of Clinical Investigation 132, doi:10.1172/jci157500 (2022).
Reimund, M. et al. Evidence for Two Distinct Binding Sites for Lipoprotein Lipase on Glycosylphosphatidylinositol-anchored High Density Lipoprotein-binding Protein 1 (GPIHBP1). Journal of Biological Chemistry 290, 13919-13934, doi:10.1074/jbc.m114.634626 (2015).
Mysling, S. et al. The acidic domain of the endothelial membrane protein GPIHBP1 stabilizes lipoprotein lipase activity by preventing unfolding of its catalytic domain. eLife 5, doi:10.7554/eLife.12095 (2016).
Kristensen, K. et al. A disordered acidic domain in GPIHBP1 harboring a sulfated tyrosine regulates lipoprotein lipase. Proceedings of the National Academy of Sciences of the United States of America 115, doi:10.1073/pnas.1806774115 (2018).
Holmes, R. & Cox, L. Comparative studies of glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1: evidence for a eutherian mammalian origin for the GPIHBP1 gene from an LY6-like gene. 3 Biotech 2, doi:10.1007/s13205-011-0026-4 (2012).
Birrane, G. et al. Structure of the lipoprotein lipase–GPIHBP1 complex that mediates plasma triglyceride hydrolysis. Proceedings of the National Academy of Sciences of the United States of America 116, doi:10.1073/pnas.1817984116 (2019).
Arora, R. et al. Structure of lipoprotein lipase in complex with GPIHBP1. Proceedings of the National Academy of Sciences of the United States of America 116, doi:10.1073/pnas.1820171116 (2019).
van Tilbeurgh H, Roussel, A., Lalouel, J. & Cambillau, C. Lipoprotein lipase. Molecular model based on the pancreatic lipase x-ray structure: consequences for heparin binding and catalysis. Journal of Biological Chemistry 269, doi:10.1016/S0021-9258(17)41822-9 (1994).
Lookene, A., Chevreuil, O., Østergaard, P. & Olivecrona, G. Interaction of Lipoprotein Lipase with Heparin Fragments and with Heparan Sulfate: Stoichiometry, Stabilization, and Kinetics. Biochemistry 35, doi:10.1021/bi960008e (1996).
Larnkjaer, A., Nykjaer, A., Olivecrona, G., Thøgersen, H. & Ostergaard, P. B. Structure of heparin fragments with high affinity for lipoprotein lipase and inhibition of lipoprotein lipase binding to alpha 2-macroglobulin-receptor/low-density-lipoprotein-receptor-related protein by heparin fragments. Biochemical Journal 307, doi:10.1042/bj3070205 (1995).
Spillmann, D., Lookene, A. & Olivecrona, G. Isolation and Characterization of Low Sulfated Heparan Sulfate Sequences with Affinity for Lipoprotein Lipase. Journal of Biological Chemistry 281, doi:10.1074/jbc.M604702200 (2006).
Young, E., Cosmi, B., Weitz, J. & Hirsh, J. Comparison of the non-specific binding of unfractionated heparin and low molecular weight heparin (Enoxaparin) to plasma proteins. Thromb Haemost 70, 625-630 (1993).
Killeen, R., Wait, R., Begum, S., Gray, E. & Mulloy, B. Identification of major heparin-binding proteins in plasma using electrophoresis and mass spectrometry. International Journal of Experimental Pathology 85, doi:10.1111/j.0959-9673.2004.390af.x (2004).
Zammit, A., Pepper, D. S. & Dawes, J. Interaction of immobilised unfractionated and LMW heparins with proteins in whole human plasma. Thromb Haemost 70, 951-958 (1993).
Xu, D. & Esko, J. D. Demystifying Heparan Sulfate–Protein Interactions. Annual review of biochemistry 83, doi:10.1146/annurev-biochem-060713-035314 (2014).
Bengtsson-Olivecrona, G. & Olivecrona, T. Phospholipase activity of milk lipoprotein lipase. Methods in enzymology 197, 345-356 (1991).
Gill, S. & von Hippel, P. Calculation of protein extinction coefficients from amino acid sequence data. Analytical Biochemistry 182, doi:10.1016/0003-2697(89)90602-7 (1989).
Damen, J., Dijkstra, J., Regts, J. & Scherphof, G. Effect of lipoprotein-free plasma on the interaction of human plasma high density lipoprotein with egg yolk phosphatidylcholine liposomes. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism 620, doi:10.1016/0005-2760(80)90188-5 (1980).
Gunn, K. H., Neher, S. B., Gunn, K. H. & Neher, S. B. Structure of dimeric lipoprotein lipase reveals a pore adjacent to the active site. Nature Communications 14, doi:10.1038/s41467-023-38243-9 (2023).
Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nature Methods 19, doi:10.1038/s41592-022-01488-1 (2022).
Evans, R. et al. Protein complex prediction with AlphaFold-Multimer. bioRxiv, doi:10.1101/2021.10.04.463034 (2022).
Consortium, T. U. et al. UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Research 51, doi:10.1093/nar/gkac1052 (2023/01/06).
Meng, E. C. et al. UCSF ChimeraX: Tools for structure building and analysis. Protein Science 32, doi:10.1002/pro.4792 (2023).
Risti, R. et al. Combined action of albumin and heparin regulates lipoprotein lipase oligomerization, stability, and ligand interactions. PLOS ONE 18, doi:10.1371/journal.pone.0283358 (2023).
Leth-Espensen, K. Z. et al. The intrinsic instability of the hydrolase domain of lipoprotein lipase facilitates its inactivation by ANGPTL4-catalyzed unfolding. Proceedings of the National Academy of Sciences 118, doi:10.1073/pnas.2026650118 (2021).
Necci, M., Piovesan, D., Predictors, C., Curators, D. & Tosatto, S. C. E. AlphaFold and Implications for Intrinsically Disordered Proteins. Journal of Molecular Biology 433, doi:10.1016/j.jmb.2021.167208 (2021).
Doolittle, R. F. in The Plasma Proteins Vol. Volume II (ed F. W Putnam) Ch. Biosynthesis Metabolism, Alterations in Disease, 148-149 (Academic Press 1975).
Olson, S. T., Halvorson, H. R. & Björk, I. Quantitative characterization of the thrombin-heparin interaction. Discrimination between specific and nonspecific binding models. Journal of Biological Chemistry 266, doi:10.1016/S0021-9258(18)38124-9 (1991).
Du, X. et al. Insights into Protein–Ligand Interactions: Mechanisms, Models, and Methods. International Journal of Molecular Sciences 17, doi:10.3390/ijms17020144 (2016).
Wallerstein, J. et al. Entropy–Entropy Compensation between the Protein, Ligand, and Solvent Degrees of Freedom Fine-Tunes Affinity in Ligand Binding to Galectin-3C. JACS Au 1, doi:10.1021/jacsau.0c00094 (2021).

No competing interests reported.

Supplementary.docx

Download PDF

Journal Publication

published 23 Aug, 2024

Read the published version in Scientific Reports →

Editorial decision: Revision requested
02 Jul, 2024
Reviews received at journal
23 May, 2024
Reviewers agreed at journal
09 May, 2024
Reviewers invited by journal
08 May, 2024
Editor assigned by journal
08 May, 2024
Editor invited by journal
08 May, 2024
Submission checks completed at journal
03 May, 2024
First submitted to journal
30 Apr, 2024

You are reading this latest preprint version

A negatively charged cluster in the disordered acidic domain of GPIHBP1 provides selectivity in the interaction with lipoprotein lipase

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Materials and methods

Reagents

Determination of LPL activity with DGGR

Modeling of the predicted complex

Results

Clusters of negatively charged residues are crucial for LPL stabilization

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1