Authors
Andres Guerrero; Stephanie Contreras; Dave Dallas; Lauren Wu; Jennifer Smilowitz; Daniela Barile; Bruce German; Carlito Lebrilla

Institutes
University of California, Davis, Davis, CA

Novel Aspect
Peptidomics analysis of milk yields bioactive peptides that vary during lactation and disease.

Introduction
Human milk, as other biofluids, contains proteolytic enzymes that produce endogenous peptides. A library of endogenous peptides found in human milk with post-translational modifications was created through the use of nano-LC and tandem mass spectrometry. This library was used to develop a method to identify peptides rapidly allowing monitoring with quantitation nearly 1000 peptides with nano-LC MS. The method was used to examine changes in the peptide abundances during lactation. A comprehensive understanding of the endogenous proteolytic activity in human milk can be achieved to compare different milk samples. In addition because many of the peptides have antimicrobial activity, we examined the effect of peptide production and abundances during diseases such as mastitis infection.

Methods
Peptide separation was performed using 25 ?L of human milk that was spiked with an internal standard. The samples were defatted and trichloroacetic acid was added to precipitate the peptides. The samples were cleaned up by solid phase extraction using a C-18 column. The samples were analyzed on a nano-LC Q-TOF with a C-18 chip to obtain MS and MS/MS data. The tandem mass spectrometry information from each sample was exported as an MGF file and uploaded to X! tandem to identify peptides present in each milk sample analyzed. When the peptides in milk were identified an in-house software was used to visualize the most abundant peptides present within a protein.

Preliminary Results/Abstract
The peptide milk library contains peptide fragments from different proteins along with their masses, retention times, and post-translational modifications that may occur on certain amino acids that include phosphorylation, oxidation, deamination, and water loss. The peptide milk library was used to identify peptides from different proteins of mothers with and without mastitis. Based on the analysis, we found that the vast majority of the peptides came from five proteins in milk: ?-casein, ?-casein, polymeric immunoglobulin, osteopontin, and butryophilin. By using an in-house software we were able to assign the peptide fragments within these proteins and identify locations where the abundances varied the most. This software also allowed us to locate the peptides within these proteins that had phosphorylation sites. For example, with ?-casein the peptides that had 1, 2, and 3 phosphorylation sites were all located at the N-terminus of the protein, whereas the peptides with no phosphorylation sites where found throughout the protein. Analysis of milk from mothers with mastitis were performed. The total peptide abundances from five proteins were summed over several time points to determine whether peptide abundances varied during mastitis. There was a general trend of a decrease in the peptide abundances around the time the mother had mastitis.

Authors
Qian Wang¹; Zachery Lewis^{2, 3}; Andres Guerrero¹; David Mills^{2, 3}; Carlito Lebrilla¹

Institutes
¹Department of Chemistry, UC Davis, Davis, CA; ²Department of Viticulture & Enology, UC Davis, Davis, CA; ³Department of Food Science & Technology, UC Davis, Davis, CA

Novel Aspect
A fast and simple method employing direct MALDI MS to compare the affinities between oligosaccharides and bacteria.

Introduction
A necessary step of bacterial colonization and infection is the adherence of the bacteria to host cell surfaces. The adherence factors of bacteria are called adhesins, which consist of two major types: protein adhesins and carbohydrate adhesins. The adherence factors of host cells which adhesins recognize are called receptors, which are usually carbohydrate or peptide residues on the surface of host cells. The specific adherence of adhesins to receptors determines the tissue specificity of binding. In this study, the goal is to develop a rapid mass spectrometry method with the use of MALDI TOF/TOF mass spectrometer to analyze the specific affinities between different bacteria and oligosaccharides which are a major type of receptors.

Methods
First, oligosaccharides solutions are applied on the MALDI spots and the MALDI plate (with immobilized phenylboronic acid on the surface which is able to bind to oligosaccharides covalently) is incubated in humidified atmosphere. After that, the unbound oligosaccharides are washed away. Then bacteria suspension is applied on the MALDI spots and the MALDI plate is incubated at 37 oC. The unbound bacteria are washed away after incubation (For the control spots, unbound bacteria are not washed away). Lastly the matrix is applied and mass spectra are obtained. If the mass spectrum matches that of the control spot, then the affinity is approved. If no (or weak) mass spectrum is observed then the affinity is disproved (or the affinity is weak).

Preliminary Data
We studied the amount of oligosaccharides to saturate the MALDI spots. If the spots are not saturated by oligosaccharides, then bacteria may bind to the plate directly which will produce false signal. Different amounts of oligosaccharides were applied on each spot and mass spectra were obtained. Starting from the spot with 10-10 mole oligosaccharides, the peak of sodiated molecular ion of the oligosaccharide was observed suggesting that the spot had been saturated. So 10-10 mole of oligosaccharides is enough to saturate a spot.

The mass spectra of the control spots were obtained. Bifidobacterium longum subsp. infantis ATCC 15697, a commensal bacteria abundant in breast fed infants, Lacto-N-tetraose, disialyllacto-N-tetraose, 3-sialyllactose, 3’-sialyllactose, 6’-sialyllactose and Lewis B tetrasaccharide were used to probe the binding interactions. Bacteria that were bound to the specific oligosaccharide-activated surfaces were analyzed by direct MALDI. The mass spectra from these compounds were the same suggesting that mass spectra were not dependent on the oligosaccharides. Escherichia Coli DH5-alpha was also applied to the surface. These bacteria yielded spectra that were distinct from B. infantis. The mass spectra of B. infantis and E. Coli with the same oligosaccharide are unique and reproducible suggesting that this method is able to differentiate different types of bacteria. An effective washing method to remove unbound bacteria was developed to probe the various bacterial surface and determine the specific affinity of the bacteria to the oligosaccharides.

Authors
Dayoung Park; Narine Arabyan; Cynthia Williams; Ting Song; Bart Weimer; Carlito Lebrilla

Institutes
University of California, Davis, Davis, CA

Novel Aspect
Methodology to analyze thousands of membrane glycan structures simultaneously with linkage and isomer differentiation and identify glycans involved in infection.

Introduction
Cell membranes consist of proteins that contain short saccharide chains called glycans, which mediate nearly all interactions. In the gastrointestinal tract, a single layer of epithelial cells lines the inner layer as a protective interface between the embedded tissues and microorganisms that continually contact the internal organs. Regulation of these bacterial-host interactions is critical for host health. Using high resolution mass spectrometric techniques, a glycomic analysis was performed to characterize glycosylation changes on epithelial cell surfaces upon prolonged contact with foreign and resident bacteria of the gut. This new technique allows for rapid profiling and quantitation of membrane glycans based on retention time and accurate mass and was used to understand which glycans are involved in bacterial colonization and infection.

Methods
Epithelial colorectal adenocarcinoma Caco-2 cells were cultured in vitro. Selected bacterial genes that express glycosidases were knocked out. Upon infection with wildtype bacteria and the knockouts, cells were lysed using probe sonication and the membrane fraction was extracted from the resulting lysate by ultracentrifugation. Glycans were enzymatically released from isolated cell membrane glycoproteins and analyzed by microfluidic chip-based nano-LC quadrupole time-of-flight mass spectrometry. The total glycan profile of detected membrane glycans were constructed using a theoretical, retrosynthetic mass library, providing glycan compositions based on known synthetic pathways. Changes in relative abundances of individual N-glycans were characterized statistically. To elucidate exact structures, tandem MS was employed and reduced glycans were digested with exoglycosidases for linkage information and isomeric differentiation.

Preliminary Results/Abstract

The cell sample set was optimized using different cell amounts varying from one to four million cells. Samples with at least 2 million cells showed the highest glycan signal and reproducibility.
Caco-2 cell surface glycosylation is dominated by sialylated and fucosylated complex and hybrid glycans, comprising almost three fourths of the total number of glycans. When considering the relative intensities, however, high mannose glycans are among the most abundant. These results suggest that Caco-2 membranes have a large amount of terminal mannose residues, which may have functional significance in epithelial cells during infection.
Bacterial infection times were varied to observe changes in Caco-2 membrane glycosylation. From 0 min to 45 min, only slight variations in glycan profiles were observed. At the 1 hr time point, there was a significant change in glycan compositions and relative abundances. Notably, nondecorated and high mannose glycans increased in signal after 1 hr of infection but later decreased past 2 hrs. By the third hour of infection, the overall profile recovered and resembled that of the uninfected set.
During the course of infection, levels of bisecting and triantennary complex glycans were significantly altered. The most abundant glycan in the uninfected sample, a bisecting, monofucosylated, bisialylated complex glycan, decreased dramatically in signal post-infection, becoming suppressed by other high abundant glycans. An isomer of this glycan, which eluted at a later time, increased in abundance fifteen fold after infection. Terminal fucose and sialic acid residues on a glycan with more than two antennas may act as receptors for bacteria and utilized as a source of energy, carbon, and nitrogen. Deficient glycan degrading enzyme activity of the bacteria led to an accumulation of certain oligosaccharide substrates on the cell surface. On average, 176 glycan compositions were identified in the uninfected sample and 166 compositions for the infected sample.

Authors
Andres Guerrero¹; Yanhong Li¹; Salem Alkanaimsh²; Lucas Arzola²; Bryce Hashimoto²; Minsook Hwang³; Aye Tu⁴; My Phu⁴; Abhaya Dandekar⁴; Bryce Falk³; Somen Nandi⁵; Raymond Rodriguez⁵; Karen McDonald²; Xi Chen¹; Carlito Lebrilla¹

Institutes
¹UC Davis, Chemistry Department, Davis, CA;²UC Davis, Chemical Engineering & Materials Science, Davis, CA; ³UC Davis, Plant Pathology, Davis, CA; ⁴UC Davis, Plant Science, Davis, CA; ⁵UC Davis, Molecular & Cellular Biology, Davis, CA

Novel Aspect:

A method for accurate glycan analysis was developed to monitor the production of a recombinant protein with multiple glycosylation sites.

Introduction:

Glycan analysis of recombinant proteins containing multiple glycosylation sites is analytically challenging especially during the initial phases of production when sample amount and purity are major limitations. In this work we propose a glycoprotein analysis method that addresses these potential limitations, provide site-specific glycan information and facilitate the quantitative comparison between samples.

The method was successfully applied during the production of a glycoengineered recombinant protein: human butyrylcholinesterase (BuChE). BuChE produced in Nicotiana benthamiana was sialytated in vitro to increase its therapeutic potential and mimic the human glycosylation pattern. We describe the glycopeptide characterization of the BuChE glycoforms expressed in different plant subcellular localizations and the glycosylation product, after every step of glycan remodeling, up to the final desired glycosylated product.

Methods:

The analytical method employed glycopeptide analysis by in-gel digestion using a non-specific protease followed by mass spectrometry. Briefly, samples were desalted by C8 solid phase extraction, denatured and run on a SDS-PAGE gel. Excised protein bands were digested with pronase E and the resulting lysates were extracted and purified for glycopeptides using graphitized solid phase extraction. Tandem-MS analysis was performed on a nanoESI-LC-Q-TOF using a graphitized chip nanoLC column. Glycopeptides were identified using GP-Finder 3 (in-house software) and the corresponding signals quantified by ion-counting. The procedure was initially optimized using another human recombinant plant-made glycoprotein, lactotransferrin produced in Oryza sativa.

Preliminary Data:

The combination of gel electrophoresis and glycopeptide analysis facilitates the analysis of specific proteins in highly impure plant extracts. Additionally, the immobilization of the glycoprotein in the gel allows analysis on even very small amounts of sample. This analytical method was successfully applied to monitor glycosylation on intermediates and products through every step of the glycoprotein production.

Substrate selection

The glycoprotein analysis showed that subcellular localization affects the N-glycan composition of the glycoprotein. The BuChE recovered from the plant endoplasmic reticulum (ER) was almost exclusively glycosylated with high-mannose, while the apoplast targeted extract (APO) was predominantly composed by the basic N-glycan plant core. Based on these results, ER BuChE would require an additional enzymatic step (trim off the high-mannose structures) to reach the desired sialylated structures compared to APO BuChE. Hence, APO BuChE was chosen as a substrate.

Method validation

To obtain the desired glycoform from plant APO BuChE, sequential addition of N-acetylglucosamine, galactose and sialic acid is required using three enzymatic systems, namely N-acetylglucosaminyltransferase, β1,4-galactosyltransferase and α2,6-sialyltransferase. The activity of these enzymes was tested using plant recombinant lactotransferrin that, similarly to APO BuChE, exhibits basic plant core N-glycosylation. The analysis showed the systematic modification of the lactotransferrin N-glycosylation pattern after each redecoration step validating both, the in vitro glycan modification and our analytical approach.

Optimization

As important as the glycan composition was to the product, the aim was product efficacy. A continuous feedback system was established between redecoration, glycoprotein analysis and protein bioactivity. APO BuChE N-glycan remodelation was performed under different conditions such as temperature and pH. At every step, product was analyzed using the analytical method to observe the progress. The final product contained at least 10%-30% of sialylation. Facilitated by our analytical method, further optimization of the redecoration process is expected to increase the degree of sialylation of the glycoprotein while keeping its activity.

Authors
Cynthia Williams; Anne Fenton; Lauren Nagy; Qiuting Hong; L. Renee Ruhaak; Satya Dandekar; Carlito Lebrilla

Institutes
UC Davis, Davis, CA

Novel Aspect
Monitoring the degree of glycosylation with the absolute protein quantitation in plasma of HIV patients.

Introduction
The immune system is highly dysregulated in patients with HIV infection. Current studies aim to understand this process to allow for better prevention, earlier diagnosis and a cure. Immunoglobulin concentrations are abnormal in HIV patients and they display altered glycosylation patterns. However, absolute protein quantitation, simultaneously, with site-specific glycosylation analysis of human plasma in HIV samples has previously not been achieved. We are developing methods for quantitation of glycosylated proteins on the protein-specific and site-specific level. In this study, we used multiple reaction monitoring (MRM) to quantitate immunoglobulins A, G, and M (IgA, IgG, and IgM) proteins and their site-specific glycans to examine whether immunoglobulin-specific glycosylation varies in HIV infection.

Methods
Plasma samples from 26 male patients and 11 controls were collected and HIV infected patients were broken into 3 groups: No Therapy (n=11), Therapy (n=11), and Long Term Non Progressors (n=4). Two uL of patient sera and IgA, IgG, and IgM standards were digested using trypsin after reduction and alkylation. Protein and glycopeptide quantitation was performed using a QqQ MS in dynamic MRM mode coupled with a UPLC system. Unglycosylated peptides were used for absolute protein quantitation. Site-specific glycosylation was normalized to absolute protein abundances to determine degree of glycosylation.

Preliminary Results/Abstract
Presented here is a high-throughput MRM method for absolute quantitation of IgG, IgA and IgM and their glycosylation analysis directly from 2 uL of serum. A 10 &11-minute UPLC gradient was used with a C18 stationary phase and were employed in the development of the method to quantify protein and site-specific glycan abundances for IgG and IgA & IgM. IgG, A, and M concentrations, as well as subclasses and glycosylation patterns were monitored through 14 MRM transitions for peptides and 53 for glycopeptides.
Our rapid throughput platform was applied to HIV samples (26 patients and 11 controls). For IgG, the peptide common from each subclass was used to quantify the total IgG content. Because this peptide is common to all four subclasses, it correlates linearly to the total IgG protein abundances. Average IgG concentrations were 10.1, 12.3, 13.7, and 19.2 mg/mL in HIV Negative (Neg), Therapy (Tx), Long Term Non-Progessors (LTNP), and No Therapy (No Tx) cases, respectively. Differences with statistical significance were observed in the protein concentrations and in specific IgG glycoforms as well. Galactose deficient glycopeptides were observed to be significantly higher in the HIV positive cases. In IgA, there were no differences observed on the protein level, but on the glycosylation level, Hex₄HexNAc₄NeuAc₁ was shown to be significantly lower in the Tx groups. In comparison to the IgG analysis, the IgM protein concentration levels were only elevated in the No Tx cases. Interestingly, although there were only 4 patient samples within the LTNP group, a positive association was observed between levels of high mannose type glycans and the LTNP response. These results show that the immune responses of HIV patients are clearly altered and that glycosylation is a very important factor.

Khaldi, N., et al. (2014). “Predicting the important enzyme players in human breast milk digestion.” J Agric Food Chem.
Human milk is known to contain several proteases, but little is known about whether these enzymes are active, which proteins they cleave and their relative contribution to milk protein digestion in vivo. We analyzed the mass spectrometry-identified protein fragments found in pooled human milk by comparing their cleavage sites with the enzyme specificity patterns of an array of enzymes. The results indicate that several enzymes are actively taking part in the digestion of human milk proteins within the mammary gland, including plasmin and/or trypsin, elastase, cathepsin D, pepsin, chymotrypsin, a glutamyl endopeptidase-like enzyme and proline endopeptidase. Two proteins were most affected by enzyme hydrolysis: beta-casein and polymeric immunoglobulin receptor. In contrast, other highly abundant milk proteins such as alpha-lactalbumin and lactoferrin appear to have undergone no proteolytic cleavage. We also show that a peptide sequence containing a known anti-microbial peptide is released in breast milk by elastase and cathepsin D.

Kim, K., et al. (2014). “Evaluation of Glycomic Profiling as a Diagnostic Biomarker for Epithelial Ovarian Cancer.” Cancer Epidemiol Biomarkers Prev.
BACKGROUND: Prior studies suggested that glycans were differentially expressed in patients with ovarian cancer and controls. We hypothesized that glycan-based biomarkers might serve as a diagnostic test for ovarian cancer and evaluated the ability of glycans to distinguish ovarian cancer cases from matched controls. METHODS: Serum samples were obtained from the tissue-banking repository of the Gynecologic Oncology Group, and included healthy female controls (n = 100), women diagnosed with low malignant potential (LMP) tumors (n = 52), and epithelial ovarian cancers (EOC) cases (n = 147). Cases and controls were matched on age at enrollment within +/-5 years. Serum samples were analyzed by glycomics analysis to detect abundance differences in glycan expression levels. A two-stage procedure was carried out for biomarker discovery and validation. Candidate classifiers of glycans that separated cases from controls were developed using a training set in the discovery phase and the classification performance of the candidate classifiers was assessed using independent test samples that were not used in discovery. RESULTS: The patterns of glycans showed discriminatory power for distinguishing EOC and LMP cases from controls. Candidate glycan-based biomarkers developed on a training set (sensitivity, 86% and specificity, 95.8% for distinguishing EOC from controls through leave-one-out cross-validation) confirmed their potential use as a detection test using an independent test set (sensitivity, 70% and specificity, 86.5%). CONCLUSION: Formal investigations of glycan biomarkers that distinguish cases and controls show great promise for an ovarian cancer diagnostic test. Further validation of a glycan-based test for detection of ovarian cancer is warranted. IMPACT: An emerging diagnostic test based on the knowledge gained from understanding the glycobiology should lead to an assay that improves sensitivity and specificity and allows for early detection of ovarian cancer. Cancer Epidemiol Biomarkers Prev; 1-11. (c)2014 AACR.

Dallas, D. C., A. Guerrero, N. Khaldi, P. A. Castillo, W. F. Martin, J. T. Smilowitz, C. L. Bevins, D. Barile, J. B. German and C. B. Lebrilla (2013). “Extensive in vivo Human Milk Peptidomics Reveals Specific Proteolysis Yielding Protective Antimicrobial Peptides.” J Proteome Res 12(5): 2295-2304.
Milk is traditionally considered an ideal source of the basic elemental nutrients required by infants. More detailed examination is revealing that milk represents a more functional ensemble of components with benefits to both infants and mothers. A comprehensive peptidomics method was developed and used to analyze human milk yielding an extensive array of protein products present in the fluid. Over 300 milk peptides were identified originating from major and many minor protein components of milk. As expected, the majority of peptides derived from beta-casein, however no peptide fragments from the major milk proteins lactoferrin, alpha-lactalbumin, and secretory immunoglobulin A were identified. Proteolysis in the mammary gland is selective-released peptides were drawn only from specific proteins and typically from only select parts of the parent sequence. A large number of the peptides showed significant sequence overlap with peptides with known antimicrobial or immunomodulatory functions. Antibacterial assays showed the milk peptide mixtures inhibited the growth of Escherichia coli and Staphylococcus aureus . The predigestion of milk proteins and the consequent release of antibacterial peptides may provide a selective advantage through evolution by protecting both the mother’s mammary gland and her nursing offspring from infection.

Ruhaak, L. R., U. T. Nguyen, C. Stroble, S. L. Taylor, A. Taguchi, S. M. Hanash, C. B. Lebrilla, K. Kim and S. Miyamoto (2013). “Enrichment strategies in glycomics based lung cancer biomarker development.” Proteomics Clin Appl.
PURPOSE: There is a need to identify better glycan biomarkers for diagnosis, early detection and treatment monitoring in lung cancer using biofluids such as blood. Biofluids are complex mixtures of proteins dominated by a few high abundance proteins that may not have specificity for lung cancer. Therefore two methods for protein enrichment were evaluated; affinity capturing of IgG and enrichment of medium abundance proteins, thus allowing us to determine which method yields the best candidate glycan biomarkers for lung cancer. EXPERIMENTAL DESIGN: N-glycans isolated from plasma samples from 20 cases of lung adenocarcinoma and 20 matched controls were analyzed using nLC-PGC-chip-TOF-MS. N-glycan profiles were obtained for five different fractions: total plasma, isolated IgG, IgG depleted plasma, and the bound and flow-through fractions of protein enrichment. RESULTS: Four glycans differed significantly (FDR<0.05) between cases and controls in whole unfractionated plasma, while four other glycans differed significantly by cancer status in the IgG fraction. No significant glycan differences were observed in the other fractions. CONCLUSIONS AND CLINICAL RELEVANCE: These results confirm that the N-glycan profile in plasma of lung cancer patients is different from healthy controls and appears to be dominated by alterations in relatively abundant proteins. This article is protected by copyright. All rights reserved.

Strum, J. S., C. C. Nwosu, S. Hua, S. R. Kronewitter, R. R. Seipert, R. J. Bachelor, H. J. An and C. B. Lebrilla (2013). “Automated Assignments of N- and O-Site Specific Glycosylation with Extensive Glycan Heterogeneity of Glycoprotein Mixtures.” Anal Chem.
Site-specific glycosylation (SSG) of glycoproteins remains a considerable challenge and limits further progress in the areas of proteomics and glycomics. Effective methods require new approaches in sample preparation, detection, and data analysis. While the field has advanced in sample preparation and detection, automated data analysis remains an important goal. A new bioinformatics approach implemented in software called GP Finder automatically distinguishes correct assignments from random matches and complements experimental techniques that are optimal for glycopeptides, including nonspecific proteolysis and high mass resolution liquid chromatography/tandem mass spectrometry (LC/MS/MS). SSG for multiple N- and O-glycosylation sites, including extensive glycan heterogeneity, was annotated for single proteins and protein mixtures with a 5% false-discovery rate, generating hundreds of nonrandom glycopeptide matches and demonstrating the proof-of-concept for a self-consistency scoring algorithm shown to be compliant with the target-decoy approach (TDA). The approach was further applied to a mixture of N-glycoproteins from unprocessed human milk and O-glycoproteins from very-low-density-lipoprotein (vLDL) particles.