Importance of glycosylation in the interaction of Tamm‐Horsfall protein with collectin‐11 and acute kidney injury

Abstract Both Tamm‐Horsfall protein (THP) and collectin‐11 (CL‐11) are important molecules in acute kidney injury (AKI). In this study, we measured the change of glycosylation of THP in patients with AKI after surgery, using MALDI‐TOF MS and lectin array analysis. The amount of high‐mannose and core fucosylation in patients with AKI were higher than those in healthy controls. In vitro study showed that THP could bind to CL‐11 with affinity at 9.41 × 10−7 mol/L and inhibited activation of complement lectin pathway. The binding affinity decreased after removal of glycans on THP. Removal of fucose completely ablated the binding between the two proteins. While removal of high‐mannose or part of the N‐glycan decreased the binding ability to 30% or 60%. The results indicated that increase of fucose on THP played an important role via complement lectin pathway in AKI.


| INTRODUC TI ON
Tamm-Horsfall protein (THP) is the most abundant protein in normal human urine, and carbohydrate forms 25% of the molecular mass of THP. 1 THP is considered as an important protective factor during acute kidney injury (AKI). 2 Lower level of urinary THP associates with increased risk of AKI in patients receiving cardiac surgery. 3 The function of THP in AKI remains unclear and the immunological activities of THP are gradually getting noticed. 4 One of the immunomodulation functions of THP is its binding capacity with complement system. 4 Previous studies revealed that Tamm-Horsfall protein could bind to complement 1q (C1q) with high affinity. 5,6 Our work also identified that THP could bind to complement factor H (cFH), and enhance the cofactor activity of cFH in degradation of C3b. 7 The interactions between THP and complements suggest that THP may be a regulating protein in complement activation. Meanwhile, complement activation gradually presents an important role in the progression of various forms of renal disease 8,9 and may also involve in AKI. 10,11 Recently, collectin-11 , an initial molecule of lectin pathway in complement activation, 12 is identified to be involved in AKI and chronic renal inflammation. 13,14 CL-11 engages with L-fucose during hypoxia-or hypothermia stress, activates the complement lectin pathway and induces renal epithelial cell injury. 13 CL-11 can also promote chronic renal inflammation and tubulointerstitial fibrosis through its effects on leukocyte chemotaxis and renal fibroblast proliferation. 14 As we mentioned above, THP can bind with C1q and cFH and participate in the regulation of classic and alternative complement pathways. Thus, it is interesting to know whether THP can also influence lectin pathway.
Collectin-11 activates lectin pathway via binding with fucose and mannose. 15 Removal of carbohydrate from THP reduced the ability of THP binding of C1q and eliminated the ability of THP to protect against complement activation. 16 Meanwhile, abnormal glycosylation of THP was observed in patients with interstitial cystitis as well as in patients accepting renal transplantation, 17,18 and the disturbance of glycosylation influenced immunoreaction of THP, indicated that glycans on THP mediated the pathogenic process of urologic diseases. 18,19 Thus, our hypothesis is that THP can bind to CL-11 and influence complement activation. The interaction is mediated by glycans on THP and change of glycosylation on THP occurred during kidney stress. We first measured the binding between THP and CL-11 and then explored whether the binding between the two molecules influenced the activation of complement system. Finally, we investigated the change of glycosylation of THP during kidney stress.

| Isolation of THP
We purified THP from urine using the rapid isolation method. 20 Briefly, urine was first filtered through diatomaceous earth in Buchner funnel; then, the earth was washed by PBS and dissolved in deionized water. The mixture was centrifuged at 20 000 g and the pellet was discarded. Sodium phosphate buffer containing 0.14 mol/L NaCl was added into the suspension. The previous steps were repeated to get the suspension again. The suspension was dialysed and the protein was concentrated with ultrafiltration cartridge (30-

| Binding of THP with CL-11 detected by microscale thermophoresis (MST)
Binding capacity of THP and CL-11 was detected by MST (Monolith NT.115, NanoTemper, Germany). THP was labelled with red fluorescent using Monolith NT™ Protein Labeling Kit. The MST power was set at 40%, and LED excitation power was set at 20%. THP labelled with fluorescent was kept at 180 nM. CL-11 was diluted from the concentration of 9 μmol/L. Measurement of the binding affinity was performed in standard treated capillaries. The data were analysed with MO Affinity Analysis software, and the binding affinity was calculated using the Hill model.

| Binding of THP with CL-11, Mannose-binding lectin (MBL) and ficolins detected by enzyme-linked immunosorbent assay (ELISA)
Binding of THP with CL-11, MBL, and ficolins was detected by ELISA. The microplates were first coated with 4μg/ml recombinant CL-11, MBL or ficolin1, 2, −3 (R&D) in 0.1 mol/L carbonate buffer (pH 9.6) at 4°C for overnight. The plates were blocked with 1% BSA/PBS at 37°C for 1 hour, and THP was added with concentration varied from 0.3125 to10 μg/mL at 37°C for 1 hour. For CL-11 and ficolins binding assay, THP was diluted in VBS buffer with 0.5 mmol/L MgCl 2 0.15 mmol/L CaCl 2 . The CaCl 2 concentration was 1 mmol/L for MBL binding assay. Subsequently, wells were incubated with rabbit anti-human uromodulin polyclonal antibody (Biomedical Technologies Inc, USA) and mouse anti-rabbit IgG alkaline phosphatase (Sigma-Aldrich, USA) both at 37°C for 1 hour.
The OD at 405 nm was measured using a microplate reader (iMark, BIO-RAD).

| Removal of N-glycans of THP
PNGaseF, endoglycosidase H (Endo H) and fucosidase were used to treat the native THP according to the product protocol. PNGaseF is an amidase which removes almost all N-linked oligosaccharides from glycoproteins. Endo H is a recombinant glycosidase which recognizes the high-mannose structure of N-glycosylation. Fucosidase is an exoglycosidase specifically recognizing fucose. N-glycan was removed both under denatured condition and native condition. In this study, denatured THP (in 100°C) lost the activity of binding to CL-11; thus, we used the recommended non-denaturing protocol to digest the THP. ELISA protocol detecting binding capacity of deglycosylated THP with CL-11 was similar as above.

| Haemolytic assay
Haemolytic assay was modified according to the previous study. 21 In this test, we detected the haemolysis percentage of chicken erythrocytes in the serum with anti-C1q antibody to exclude the interference of classical complement pathway. 21

| Study participants
Twenty patients in intensive care unit (ICU) were recruited in this study. The 20 individuals were admitted into ICU because of various type of surgery, and they were diagnosed as acute kidney injury on an increase in serum creatinine ≥0.3 mg/dL (26.5 μmol/L) within 48 hours or urine volume <0.5 mL/kg/h for 6 hours, based on KDIGO recommendation. 22 Their characteristics were listed in Table 1. Informed consents were obtained from participants. Urine of all the participants was collected, and Tamm-Horsfall protein was purified and prepared for further detection of glycans by matrixassisted laser desorption/ionization-time of flight mass spectrometer (MALDI-TOF MS) and lectin array analysis. We also purified THP from 10 individuals with normal renal function as controls in the analysis.

| Measurement of N-glycan profile by MALDI-TOF MS
One hundred microgram THP was used in this experiment.
N-glycan of THP was released by PNGase F (Sigma) after reduction and alkylation. THP was first denatured at 100°C for 10 minutes.
And then the protein was reduced in 20 mmol/L dithiothreitol at 56°C for 45 minutes, followed by alkylated with iodoacetamide

TA B L E 1 Characteristics of the participants
Note: The characteristics of patients were listed above. The patients were diagnosed with AKI based on (1) an increase in serum creatinine ≥ 0.3 mg/ dL (26.5 μmol/L) within 48 h or (2) urine volume <0.5 mL/kg/h for 6 h, from KDIGO recommendation. The stage of AKI was also identified according to KDIGO recommendation. 22 Data were expressed as mean ± SEM.

| Statistical analyses
All data were analysed using SPSS 22.0 (IBM, USA

| THP bound with collectin-11 in a dosedependent manner
The purity of collected Tamm-Horsfall protein was measured by Coomasie blue stain ( Figure S1). The binding capacity of THP with CL-11 was detected by ELISA and MST. The result showed that THP bound to CL-11 in a dose-dependent manner ( Figure 1A). The binding capacity was re-confirmed by MST, with a binding affinity at 9.41 × 10 −7 mol/L ( Figure 1B).

| Glycans on THP mediated binding between THP and collectin-11
PNGaseF was used to remove the carbohydrate side chains in non-  Figure 1C).

| THP inhibited erythrocytes haemolysis mediated by lectin pathway
Haemolytic assay was a typical functional study to test complement activity. Activation of each of the three complement pathways can induce haemolysis. We first set up a system to test the activation of lectin pathway, which was modified according to previous study. 21 In mixture without THP, the haemolysis was nearly 100% and adding of THP could inhibit haemolysis of erythrocytes. Haemolysis inhibition ratio of THP was dose dependent when pre-incubated THP and serum in 37°C. At the concentration of 10 μmol/L, THP could inhibit nearly 90% complement activity.
When pre-incubated BSA instead of THP, there was no inhibition of haemolysis ( Figure 2).
In order to identify which components in lectin pathway play crucial roles in haemolysis inhibition test. We also detected THP  26 Binding capacity of THP with ficolin1 and ficolin2 was stronger than ficolin3 ( Figure 3).

| Change of N-glycan spectra in patients with acute kidney injury, by MALDI-TOF-MS
Ten patients were confirmed with newly onset acute kidney injury after surgery (  Figure 4A,B).

| Change of N-glycans in patients with acute kidney disease, by lectin array
We also measured the N-glycan changes using lectin array, as verification of MALDI-TOF-MS findings in 15 AKI patients (Table 1) Raw data were normalized first, and independent sample t test was used to compare the difference between AKI patients and healthy controls. *P < .05, **P < .01. False discovery rate was used to correct the P-values for multiple testing (Table S2) S2A displayed a representative picture of the profiling of lectin array and  (Table 2).
Other lectins recognizing fucose such as Lotus, RS-Fuc, AAL had a trend of elevation in AKI although there were no significant differences (Table   S1). There was no difference of sialylation in AKI patients and controls.

| Functional assays on the isolated THP from human AKI
The binding of CL-11 to THP isolated from AKI patients and healthy controls was also measured by ELISA. Binding strength of CL-11 to THP collected from AKI patients was higher when compared to normal controls especially at 2.5 µg/mL of THP ( Figure 5A). The haemolysis assay also showed that THP from AKI patients had enhanced inhibition on haemolysis of red blood cells, especially at the concentration of 0.3125 µmol/L ( Figure 5B).

| D ISCUSS I ON
In this study, we identified a binding capacity between THP and CL-11. The binding affinity was at 9.41 × 10 −7 mol/L, lower than antigen-antibody interaction, 27   CL-11 appeared to direct the innate immune response against locally invasive factors 33 following interaction with stress-induced l-fucose in renal epithelial cells. 13 Thus, we used different enzymes to remove various glycans on the surface of THP. PNGaseF can remove all carbohydrate side chains without selection, but in non-reduced condition, the 3-dimensional structure of protein will influence the digestion of glycan chains. Fucosidase and EndoH are enzymes that can remove fucose and high-mannose glycans specifically. In our study, binding between THP and CL-11 was decreased with removal of different type of glycans, such as fucose, mannose or complex structure from THP.
THP is a glycoprotein with core fucosylation. 30 Removal of fucose with fucosidase completely ablated the binding between CL-11 and THP, indicating that fucose on THP was crucial factor mediating the binding. Since the CRD region of CL-11 was the major structure to recognize mannose or fucose, and the binding affinity was significantly decreased when removing the glycans from THP, we speculated that THP bound with CL-11 via its CRD region.
Because the interaction of THP and CL-11 is mediated by glycans on THP, it is interesting to know whether abnormal glycosyla- lectin array analysis showed that GalNAc was the major declined complex structure in AKI patients. In this study, we found that glycans with high-mannose or fucosylation were higher and glycans with more complex structure were lower in patients with AKI compare with controls. This phenomenon may be explained by the sequential glycosidase and glycosyltransferase reactions during N-glycosylation.
Usually the N-glycan is first trimmed by glucosidases and mannosidases to Man5GlcNAc2, which serves as the precursor to further complicated glycan structures. 34 Glycans that are not processed or incompletely processed by the mannosidases are named high-mannose glycans. Fucosyltransferase eight catalyses the addition of α1,6-linked fucose residues to the first GlcNAc residue, named as core fucosylation. 34 High-mannose glycans and fucosylation are early stage of glycosylation and usually with low-molecular weight. More complicated poly-antennary structures are generated based on early stage. 34,35 The more complex structure of glycans usually have high molecular weight. During tissue injury, the glycosylation procedure may be disrupted at different stage, 18 thus the more complex structures tend to decrease. Similar phenomenon was also reported in another study. 36 We then did a binding analysis and haemolytic analysis to know is whether glycosylation change is a common phenomenon of AKI compared with normal status, we did not limit the type of surgery and selected controls with normal renal function without surgery. In the lectin microarray analysis, although only AAA reached significant p value, the other lectins also present the same trend. The limitation of sample size may be one of the reasons, except for tissue or protein specificity. Although the haemolysis assay was modified, influence of C1q and MBL were excluded, ficollins and other colletin molecules cannot be completely excluded from this system.
However, the structure and carbohydrate ligand specificity of them are different. Ficolins recognize N-acetylated molecule 43 which is not obviously involved in AKI.
In summary, THP can bind to CL-11, one of initiated molecules of lectin pathway and inhibit complement system activation via the lectin pathway. The interaction of THP and CL-11 is mainly mediated by fucoses of THP. Higher level of fucose and mannose of THP could be found in acute kidney injury patients. We propose that this interaction between the lectin pathway and THP may provide a new insight for the immunomodulatory activities of THP.

ACK N OWLED G EM ENTS
We are grateful to the volunteer who participated in this study.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
Yuqing Chen conceived and designed the study. Kunjing Gong, Min

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.