Protein composition analysis of human plasma-derived and recombinant human serum albumin preparations based on 4D label-free proteomics

HSA preparations

Two rHSA and six pHSA 20% preparations were donated by eight blood product companies in China: yeast-derived rHSA (YrHSA) from A company, oryza sativa derived rHSA (OsrHSA) from B company and pHSAs from C, D, E, F, G and H companies.

Sample preparation

Samples were removed from −80 °C, and an inhibitor (1% protease inhibitor) was added. The sample was sonicated and lysed. After centrifugation at 12,000 g and 4 °C for 10 min, the supernatant containing soluble proteins was carefully transferred to a fresh centrifuge tube. The molecular weight of proteins in each sample was analyzed by SDS-PAGE. According to the protein concentration measurement results, an equal amount of protein was taken for each sample and added to a centrifuge tube. 5 µL of 4× Loading buffer was added, then the total volume 20 µL was achieved by addition of 2% SDS. Loading: 1 µL pre-stained protein marker and 20 µL protein samples were loaded sequentially. The blank control was filled with 20 µL of 1× Loading buffer in the adjacent empty wells. Electrophoresis: the proteins were concentrated into a single line in the gel with15 mA/gel, approximately 15 min. Then, the separation gel was run at 35 mA until the dye reached the bottom of the gel. Staining and destaining: the gel was stained in Coomassie Brilliant Blue R-250 staining solution at room temperature for 2 h. Then, deatained by destaining solution and destained until the background became colorless and the bands became clear. Each sample was digested with same amounts, and then adjusted to the same volume with lysis solution (8 M urea). The samples were reduced using 5 mM dithiothreitol (30 min at 56 °C) and alkylated with 11 mM iodoacetamide (15 min at room temperature in darkness). After that, the urea concentration was adjusted to less than two M by adding 200 mM TEAB. Trypsin was added at 1:50 trypsin-to-protein mass ratio for an initial digestion overnight at 37 °C and 1:100 trypsin-to-protein mass ratio for a second 4 h digestion at 37 °C. At last, a Strata X SPE column was used to desalt the peptides.

C18 Tip fractionation

In order to separate the peptides into several fractions, a homemade tip column (C18, 3 µm, 150 Å, Agela) was used. First, was used to activate The tip column was activated by 50 µL 100% acetonitrile (ACN). Then, 50 µL 50% ACN and 50 µL phase A (98% H₂O, 2% ACN, 0.1% formic acid) were used to equilibrate the tip column. The peptides were loaded into the tip twice in 50 µL phase A, and washed with phase A. Finally, the peptides were eluted into six fractions by increasing the ACN gradient: 7%, 10%, 17%, 23%, 28% and 50%. Then the fractions were combined into four fractions, Fra1 (7%, 28%), Fra2 (10%, 50%), Fra3 (17%) and Fra4 (23%), and dried by vacuum centrifugation.

LC-MS/MS analysis

After dissolved in solvent A, the tryptic peptides were directly loaded onto a homemade reversed-phase analytical column (100 µm i.d. × 25 cm) packed with 1.9 µm/120 ÅReproSil-PurC18 resins (Dr. Maisch GmbH, Ammerbuch, Germany). Solvent A (0.1% formic acid, 2% ACN in water) and solvent B (0.1% formic acid, 90% ACN in water) together made the mobile phase. The peptides were then separated with the following gradient: 0–68 min, 6%–23% B; 68–82 min, 23%–32% B; 82–86 min, 32%–80% B; 86–90 min, 80% B. An EASY-nLC 1,200 UPLC system from ThermoFisher Scientific was used with a constant flow rate of 500 nl/min. Then an Orbitrap Exploris 480 with a nano-electrospray ion source was used to analyze the separated peptides . And the FAIMS compensate voltage (CV) used was −45 V and −65 V. The Orbitrap detector was used to analyze precursors and fragments. The full MS scan resolution was set to 60,000 for a scan range of 400–1,200 m/z. 25 of the most abundant precursors were selected for further MS/MS analysis with 20 s dynamic exclusion, and the MS/MS scan was fixed as 110 m/z at a resolution of 15,000 with Turbot MT The HCD fragmentation was performed at a normalized collision energy of 27%. The automatic gain control target was set at 100%, with an intensity threshold of 50,000 ions/s and Auto maximum injection time.

Data analysis and bioinformatics

MaxQuant search engine (v.1.6.15.0) was used to process the resulting MS/MS data. Tandem mass spectra were searched against the UniProt database (http://www.uniprot.org) (Homo_sapiens_9606_SP_20230103.fasta, Oryza_sativa_subsp.japonica39947_PR_20221125. fasta and Saccharomyces_cerevisiae_4932_UP_20230316_seqkit.fasta) concatenating with the reverse decoy and contaminants database. The cleavage enzyme used was Trypsin/P, with two missing cleavages allowed. The minimum peptide length was set to seven, with a maximum of five modifications per peptide. The mass error tolerance was set as 20 ppm for precursor ions both in the first search and main search.

GO annotation was performed to annotate and analyze the identified APs with eggnog-mapper software (v2.0). The software is based on the EggNOG database (http://eggnog5.embl.de/#/app/home). The GO ID was extracted from the results of each protein annotation, and then the protein was classified according to cellular component, molecular function and biological process. The protein domains of the identified APs were annotated based on the Pfam database (https://www.ebi.ac.uk/interpro/entry/pfam/#table) and the corresponding PfamScan tool. WolF PSORT software was used to predict the subcellular localization of all identified APs. The NCBI COG (https://ftp.ncbi.nih.gov/pub/COG/) and EggNOG (http://eggnog5.embl.de/#/app/home) databases were used to provide a more comprehensive classification of species and protein sequences and to facilitate phylogenetic tree construction and functional annotation of each homologous gene cluster. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.kegg.jp/kegg/mapper.html) was used for KEGG pathway enrichment analysis. Fisher’s exact test was used to analyze the significance of KEGG pathway enrichment for the identified APs (using all proteins in the species database as the background), and P values <0.05 were considered significant. The Reactome database (https://reactome.org/) and the WiKipathways database (https://www.wikipathways.org/) were only used for the pathway enrichment analysis of the APs identified in the pHSA preparations.

Validation of proteomics results

A validation set (including the six pHSA samples) was detected by ELISA and according to the manufacturer’s protocols to verify the results of 4D label-free-based quantitative proteomics. The candidate APs validated by ELISA were haptoglobin (HP), hemopexin (HPX), serotransferrin (TF), transthyretin (TTR) and apolipoprotein M (ApoM). The selection of five proteins for ELISA validation was based on: (1) potential functional significance in the human body; (2) different functional categories or pathways; and (3) availability in commercial ELISA kits. The ELISA kits of HP, TF and TTR were purchased from Elabscience (Wuhan, China), and the HPX and ApoM ELISA kits were from FineTest (Wuhan, China).

Molecular weight of proteins in HSA preparations

The SDS-PAGE analysis of proteins in HSA preparations revealed that the primary component, the albumin band was consistent with the expected molecular weight range of approximately 66.5 kDa (Fig. 1). Furthermore, the results demonstrated the presence of various other proteins with different molecular weights in all HSA preparations. Intriguingly, most APs exhibited molecular weights greater than albumin. Additionally, the pHSA preparations displayed a higher abundance and intensity of accompanying protein bands than the rHSA preparations.

Identification of proteins in HSA preparations

The validation of MS data is shown in Figs. S1–S8. For all eight HSA samples, most of the peptides exhibited a distribution ranging from 7 to 20 amino acids, aligning with the general rule of fragmentation based on enzymatic hydrolysis and MS. The distribution of peptide lengths identified by MS met the quality control requirements. The coverage of most proteins was below 30%. In the shotgun (also called bottom-up) strategy, MS preferentially scans peptides that are more abundant. Consequently, there is a positive correlation between protein coverage and abundance in samples.

A total of 456 APs in all pHSA samples (company C: 186; D: 190; E: 166; F: 148; G: 342; and H: 182 Aps) were identified, whereas 52 APs in the YrHSA (yeast derived rHSA) sample and 152 in the OsrHSA were identified by 4D-LFQ proteomics (Tables S1–S8). The molecular weights of these APs identified in HSA preparations from three different sources are shown in Fig. 2. The frequency of 20–30, 10–20 and 50–70 kDa APs was most pronounced in the YrHSA, OsrHSA and pHSA samples after eliminating duplicates when merging data from the six pHSA samples, respectively.

The relative abundance and protein score of all APs in each sample were calculated (Table S9). According to the relative abundance value of each AP, three proteins (84.6% K7_Dit1p, 8.7% DASH complex subunit ASK1, 5.1% Nucleolar protein 9) in YrHSA and eight proteins (63.3% Cytochrome c oxidase subunit 5C, 11.7% LOC_Os10g39430, 8.0% Terpene cyclase, 2.3% Os11g0303200, 1.9% Superoxide dismutase (Cu-Zn) chloroplastic, 1.6% Superoxide dismutase (Cu-Zn) 1: SODCC1, 1.1% Lactoylglutathione lyase and 1.0% Early nodulin-like protein 1) in OsrHSA had a relative abundance value greater than 1% (Fig. 3). Among the six pHSA samples, the eight proteins with the highest relative abundance in each pHSA sample included HP, HPX, TF, TTR, plasma protease C1 inhibitor (SERPING1), alpha-1-acid glycoprotein 1 (ORM1), keratin type I cytoskeletal 9 (KRT9), keratin type II cytoskeletal 1 (KRT1), hemicentin-1 (HMCN1), beta-2-glycoprotein 1 (APOH), apolipoprotein A-II (APOA2), alpha-2-HS-glycoprotein (AHSG), afamin (AFM) and alpha-1B-glycoprotein (A1BG). Notably, the relative abundance of HP, HPX and TTR in all pHSA samples exhibited relatively high levels (Fig. 4).

All identified APs in the pHSA samples were analyzed by UpSetR (Fig. 5). Ninety-six APs were identified in the six pHSA samples. Additionally, 156 proteins were identified exclusively in the pHSA sample from G company. Furthermore, 22, 19, 15, 12 and 11 proteins were identified exclusively in the pHSA samples from D, H, C, E and F companies, respectively.

Bioinformatics analysis of APs in HSA

To comprehensively understand the functional properties of the identified APs (merging data from the six pHSA samples and removing duplicates) in pHSA, various aspects of these proteins were annotated, including Gene Ontology (GO), protein domain, KEGG pathway, COG/KOG functional classification, subcellular localization, Reactome and WikiPathways (Table S10). Enrichment analysis of the identified APs was then performed (note: bioinformatics analysis of APs identified in rHSA preparations are available to examine in Files S1 and S2).

According to Gene Ontology, the APs in pHSA clustered primarily into 43 GO functional categories, including 17 biological processes, 12 cellular components and 14 molecular functions (Fig. S9). Subcellular localization analysis showed that these proteins in pHSA samples mainly included 132 extracellular proteins, 111 cytoplasm-related proteins, 92 nucleus-related proteins, 39 plasma membrane-related proteins, 32 mitochondria-related proteins, 18 cytoplasm nucleus-related proteins, 18 endoplasmic reticulum-related proteins and 14 other undefined proteins (Fig. S10). Functional classification of these APs was carried out using the COG/KOG protein databases. The results revealed that the APs in pHSA were classified into 24 COG/KOG categories, among which “posttranslational modification, protein turnover, chaperones” represented the largest group and involved 51 proteins (Fig. S11). KEGG pathway analysis showed that these proteins were involved in 42 pathways: four cellular processes, three environmental information processing, three genetic information processing, 11 human diseases, 11 metabolism and 11 organismal systems (Fig. S12).

GO enrichment analysis revealed that the most significantly enriched biological processes were the regulation of proteolysis, negative regulation of endopeptidase activity and negative regulation of peptidase activity. The top two cellular components were enriched in the extracellular region and extracellular space. The main molecular functions were endopeptidase inhibitor activity, peptidase inhibitor activity and endopeptidase regulator activity (Fig. 6 and Table S11).

Enrichment analysis of the protein domains in the APs revealed that the most enriched term was serpin (serine protease inhibitor) (Fig. 7A and Table S12). Most of the significantly enriched domains are directly related to the regulation of serine and other proteases that play roles in coagulation, fibrinolysis, inflammation, wound healing and tissue repair. The enrichment analysis of the KEGG pathway (Fig. 7B and Table S13) revealed that the top four enriched pathways were complement and coagulation cascades, Staphylococcus aureus infection, cholesterol metabolism and the estrogen signaling pathway. Moreover, 19, 19, 10 and 16 APs, respectively, were identified to be involved in the aforementioned pathways. The pathway enrichment results showed that these APs were involved in three primary Reactome pathways: neutrophil degranulation, innate immune system and immune system (Fig. 7C and Table S14). Furthermore, enrichment analysis revealed that the most significantly enriched WikiPathway was complement and coagulation cascades, and the next enriched pathways were the network map of the sarscov2 signaling pathway and aerobic glycolysis (Fig. 7D and Table S15).

Protein–protein interaction network analysis of APs and albumin in pHSA

We constructed a PPI proteomic network using the STRING database to understand better the interactions among the APs in pHSA and albumin. PPI analysis revealed that 49 APs interacted with albumin, of which 24 might be upstream interacting proteins and 25might be downstream interacting proteins (Tables S16 and S17). The robust and crosstalk signaling cascades are mapped in Fig. 8. Additionally, two closely interacting protein clusters were screened by the MCODE method. KEGG pathway enrichment analysis was performed on these two protein clusters, and the top two significantly enriched pathways was the biosynthesis of amino acids and cholesterol metabolism.

Validation of proteomics results

ELISA detection was performed for five proteins, i.e., HP, HPX, TF, TTR and ApoM, to validate the quantitative data obtained by 4D label-free quantitative proteomics. The ELISA results revealed that HP, HPX, TF and TTR were detectable in pHSAs but not in rHSAs, which is consistent with the results of 4D label-free quantitative proteomics (Fig. 9). However, because of the extremely low levels of ApoM and TF, it was challenging to reach the detection limit of the ELISA assay kit. TF was not detected in the pHSA sample from G company (Fig. 9D), and ApoM was not detected in all pHSA samples by ELISA. Furthermore, the ELISA results indicated significant differences in HP, HPX, TF and TTR levels among pHSA samples obtained from different manufacturers.