Protein recycling in Bering Sea Algal incubations
Protein present in phytoplankton represents a large fraction of the organic nitrogen and carbon transported from its synthesis in surface waters to marine sediments. Yet relatively little is known about the longevity of identifiable protein in situ, or the potential modifications to proteins that occur during bloom termination, protein recycling and degradation. To address this knowledge gap, diatom-dominated phytoplankton was collected during the Bering Sea spring blooms of 2009 and 2010, and incubated under darkness in separate shipboard degradation ex periments spanning 11 and 53 d, respectively. In each experiment, the protein distribution was monited over time using shotgun proteomics, along with total hydrolyzable amino acids (THAAs), total protein, particulate organic carbon (POC) and nitrogen (PN), and bacterial cell abundance. Identifiable proteins, total protein and THAAs were rapidly lost during the first 5 d of enclosure in darkness in both incubations. Thereafter the loss rate was slower, and it declined further after 22 d. The initial loss of identifiable biosynthetic, glycolysis, metabolism and translation proteins after 12 h may represent shutdown of cellular activity among algal cells. Additional peptides with glycan modifications were identified in early incubation time points, suggesting that such protein modifications could be used as a marker for internal recycling processes and possibly cell death. Protein recycling was not uniform, with a subset of algal proteins including fucoxanthin chlorophyll binding proteins and RuBisCO identified after 53 d of degradation. Non-metric multidimensional scaling was used to compare the incubations with previous environmental results. The results confirmed recent observations that some fraction of algal proteins can survive water column recycling and undergo transport to marine sediments, thus contributing organic nitrogen to the benthos.
Sample Processing Protocol
"Bering Sea water was collected during Bering Sea Ecosystem Study (BEST; Wiese et al. 2012) cruises on the outer shelf during the spring of 2009 (59.9037°N, 176.1278°W; sampling depth 5 m; water column depth 136 m) and 2010 (56.7272° N, 169.4271°W; sampling depth 36 m; water column depth 104 m) in areas which coincided with the developing spring bloom adjacent to the retreating ice. The phytoplankton community at the time of sampling was diatom dominated (Lomas et al. 2012, Moran et al. 2012). In each year, single 20 l carboys were filled from the CTD rosette taken from the chlorophyll maximum based on chlorophyll fluorescence at the time of sampling. In order to increase the amount of algal material for analysis throughout the incubation, 1 l of concentrated phytoplankton material was obtained by gently passing 10 l of CTD water from an additional bottle through a 10 µm mesh and combined with untreated seawater to make up the 20 l incubation. Macrozooplankton were excluded from the incubation by passing CTD water through a 1 mm plankton net before being added to the 20 l carboys. Incubations were placed in shipboard −1°C cold rooms for 11 d (2009) and 53 d (2010) for the duration of the experimental period. Carboys were covered to exclude light throughout the incubations and aerated with filtered air. At regular time points, carboys were gently mixed until algal material was homogeneous - ly distributed, and 1 l water samples were collected and filtered onto 25 mm pre-combusted glass fiber filters (GF/F) and 37 mm polycarbonate (0.2 µm) filters for analysis (Table 1). In addition, whole water samples were collected at each time point and filtered onto 0.2 µm filters, DAPI stained, and fixed onto microscope slides for bacterial counts. All incubation particles and bacterial slides were stored at −70°C until analysis or counting. Stained bacterial cells were counted on an Olympus BH2-RFCA fluorescent microscope. Total hydrolyzable amino acids (THAAs) were identified and quantified by gas chromatography/ mass spectrometry (GC/MS) using the EZFaast method (Phenomonex), which uses derivatization of amino acids with propyl chloroformate and propanol for detection (Waldhier et al. 2010). Briefly, suspended particles collected on GF/Fs were hydrolyzed for 4 h at 110°C (Cowie & Hedges 1992) with 6 M ana lyticalgrade HCl and L-γ-Methylleucine as the recovery standard. Following hydrolyzis and derivatization, amino acids were quantified using an Agilent 6890 capillary GC with samples injected at 250°C and separated on a DB-5MS (0.25 mm ID, 30 m) column with H2 as the carrier gas. The oven was ramped from an initial temperature of 110°C to 280°C at 10°C min−1 followed by a 5 min hold. Amino acid identification was accomplished by an Agilent 5973N mass spectrometer run under the same conditions with helium as the carrier gas and mass spectral acquisition over the 50 to 600 Da range. The protein bovine serum albumin (BSA) was analyzed in parallel to correct for responses among individual amino acids and calculation of molar ratios. The analytical precision (% relative standard deviation) for amino acid analysis was ±5%. Amino acids were normalized to percent carbon and nitrogen present in bulk samples analyzed by standard combustion methods. Total protein content was also estimated by the Bradford assay (Bradford 1976). Incubation particles collected on polycarbonate filters were extracted for proteins with pulse sonication in 6 M urea with a Branson 250 sonication probe at 20 kHz for 30 s on ice. The extracts were then frozen at −80°C, thawed, and sonicated again for 30 s on ice. This was repeated for a total of 5 sonications and 4 freeze/thaw cycles. Filter extracts of each incubation time point were then digested in 3 replicate groups: (1) standard tryptic digestion with reduction and alkylation (Nunn et al. 2010); (2) digestion with Endoproteinase GluC (Endo GluC), which cleaves peptide bonds C-terminal to glutamic acid (Drapeau et al. 1972) and to a lesser extent aspartic acid (Birktoft & Breddam 1994), to increase the number of proteins identified; and (3), incubation with Peptide N-Glycosidase F (PNGase F), which hydrolyzes nearly all types of N-glycan chains from glycoproteins and glycopeptides (Maley et al. 1989), in order to observe potentially modified proteins prior to tryptic digestion. All digests were concentrated using a speedvac to a volume that gave a final protein concentration of 1 µg per 10 µl based on measured protein concentrations of filter extracts. The uniform 1 µg per 10 µl protein concentration ensured that results would not be biased by sampling, or protein concentration differences at different incubation time points."
Data Processing Protocol
"Protein identification of sample digests was performed via shotgun proteomic tandem mass spectrometry (MS2; Aebersold & Mann 2003). Digests were analyzed using full scan (specific mass to charge ratio [m/z] 350−2000), followed by gas phase fractionation with repeat analyses over multiple narrow, but overlapping, m/z ranges (Yi et al. 2002, Nunn et al. 2006). Mass spectra were evaluated and database searched with an in-house copy of SEQUEST (Eng et al. 1994, 2008). All searches were performed with no assumption of proteolytic enzyme cleavage (e.g. trypsin, Endo GluC) to allow for identification of protein degradation products due to microbial recycling. A fixed modification was set for 57 Da on cysteine and a variable modification of 16 Da on methionine resulting from alkylation and reduction steps, respectively. A variable 1 Da modification was set for asparagine on PNGase F + trypsin digested samples to account for the conversion of asparagine to aspartic acid after cleavage of glycan chains with the use of PNGase F (Plummer et al. 1984), which takes place specifically at the consensus sequence Asn-XxxSer/Thr where Xxx can be any amino acid except proline (Bause & Hettkamp 1979). Each tandem mass spectrum generated was searched against a protein sequence database to correlate predicted peptide fragmentation patterns with observed sample ions. Probabilistic scoring of protein identifications employed PeptideProphet and ProteinProphet (Keller et al. 2002, Nesvizhskii et al. 2003) with thresholds set at 90% confidence levels on PeptideProphet and ProteinProphet for positive protein identifications from SEQUEST search results. Mass spectra from all samples were searched against a database containing the proteomes of Thalassiosira pseudonana (marine diatom), Prochlorococcus marinus (marine cyanobacterium), and Candidatus Pelagibacter ubique (marine bacterium belonging to the SAR11 clade). These proteomes were selected to follow protein degradation in a diatom dominated system with potential input of bacterial proteins (e.g. Nunn et al. 2010). Results of database comparison studies showed functional agreement for over 95% of identified peptides between the T. pseudonana - P. marinus - Ca. P. ubique database versus the larger NCBI non-redundant database containing over 11 million protein sequences (Moore et al. 2012b). To better compare shipboard incubations with previous field observations, non-metric multidimensional scaling (NMDS, theory and applications described in Borg & Groenen 2005), was performed using R statistical software to group incubation time points with Bering Sea water column particles and surface sediments from a previous study (Moore et al. 2012a), based on the distribution of identified proteins in each sample. Suspended water column particles, sinking sediment trap material, and surface sediments analyzed by Moore et al. (2012a) were collected before, during, and after the spring 2009 phytoplankton bloom and analyzed using the same proteomic methods as samples from experimental incubations."