Shedding light on cobalamin photodegradation in the ocean

Cobalamin, vitamin B12, is an important micronutrient that has been investigated for decades in the marine context because it is required for phytoplankton growth. The biologically active forms (Me‐B12, Ado‐B12) and the synthetic form (CN‐B12) quickly convert to OH‐B12 after light exposure in various aqueous solutions, but puzzlingly have been frequently reported to dominate dissolved cobalamin pools in the sunlit ocean. Here, we document photodegradation timescales for these cobalamin forms in natural seawater using targeted mass spectrometry, providing quantitative evidence that OH‐B12 is expected to be the dominant dissolved form in irradiated seawater. Then, through high‐resolution mass spectrometry, we identify four photodegradation products of OH‐B12 which represent potential building blocks microbes could salvage and remodel to satisfy cellular cobalamin requirements. Taken together, these results clarify the impact of light on marine cobalamin dynamics, laying a foundation for a more quantitative understanding of the role of cobalamin in microbial communities and biogeochemical cycles.

Cobalamin is a large, cobalt-containing corrin ring coordinated to 5,6-dimethylbenzimidizole (DMB) (alpha ligand) and either cyano-(CN), hydroxyl-(OH), methyl-(Me), or adenosyl-(Ado) (beta ligands) (Fig. 1A).Me-and Ado-B 12 are non-interchangeable enzymatic co-factors (Banerjee and Ragsdale 2003).CN-B 12 is a product of synthetic cobalamin synthesis, not known to be naturally produced (Roth et al. 1996;Warren et al. 2002) and OH-B 12 is a reaction intermediate.OH-B 12 and CN-B 12 are not biologically active and are converted to Me-or Ado-B 12 for use by cells.Ado-, Me-, and CN-B 12 convert to OH-B 12 in seconds to minutes of light exposure in blood, plasma, and slightly acidified water (Juzeniene and Nizauskaite 2013;Vaid et al. 2018;Möller et al. 2022).From this, it is expected that, in the presence of light, OH-B 12 would be the dominant form in aqueous solutions, including seawater.However, the sole examination of cobalamin degradation in natural seawater was conducted in the 1960s via bioassay (Carlucci and Silbernagel 1969).S1.
Due to analytical limitations at the time, the photolability of specific cobalamin forms was not investigated.
To date, only 26 studies have published dissolved cobalamin measurements in the ocean (Fig. 1B,C).Such measurements began in the 1950s using bioassays, whereby concentrations were inferred by measuring growth of auxotrophic organisms resulting from addition of the sample (Supporting Information Table S1).These bioassay studies report total cobalamin concentrations similar to those obtained using modern methods (Fig. 1C).Modern liquid chromatography-mass spectrometry (LC-MS) approaches now have the ability to resolve different cobalamin forms.However, there are large inconsistencies in the dominant forms of cobalamin reported in the sunlit surface ocean (Fig. 1B).For example, Heal et al. (2014) determined that OH-B 12 was most abundant in the water column in Hood Canal (#18) while Suffridge et al. (2017) suggested that Me-B 12 was consistently dominant in the Eastern Atlantic Ocean (#19).Our current understanding of the photolability of dissolved Ado-B 12 and Me-B 12 in aqueous solutions suggest its dominance in the surface ocean would be surprising.This lability has, however, not yet been quantified in seawater.
Cobalamin degradation products were investigated contemporaneously with the structural elucidation of the molecule during the 1950s with a focus on products generated from acid hydrolysis (Brink et al. 1950;Kuehl et al. 1955) but, to our knowledge, photodegradation products have yet to be investigated in marine systems or using modern approaches for characterization.There are two primary reasons that cobalamin degradation products may be of interest in marine systems: (1) they may constitute some of the poorly characterized cobalt-binding ligands which play important roles in the cobalt cycle (Saito et al. 2004;Wong et al. 2022) and (2) they may satisfy cobalamin demand in certain microbes through salvage and remodeling.Our understanding of the potential role of cobalamin degradation products lags behind other B-vitamins.For example, different microbes are able to meet thiamine (B 1 ) requirements using a range of specific precursors and degradation products (Carini et al. 2014;Paerl et al. 2015Paerl et al. , 2023)).Although we know that the genetic capacity to salvage degraded B 12 is present in select bacteria, the structures and sources (biotic or abiotic) of specific molecules remain unclear (Shelton et al. 2019).Identifying photodegradation products is a key step in understanding the role they play in biogeochemical cycles.
In many ways, cobalamin is one of the most well-studied marine metabolites.However, substantial knowledge gaps impede our understanding of its role in marine ecosystems.
Here we work to address some of these gaps through measurement of photodegradation of Me-, Ado-, CN-, and OH-B 12 in a natural seawater matrix to quantify cobalamin form-specific timescales of photodegradation.Then, through highresolution mass spectrometry, we identify OH-B 12 photodegradation products.These results can be leveraged to evaluate analytical approaches and prioritize detection of the important cobalamin forms and degradation products in the ocean.Furthermore, similar approaches can be used to investigate the influence other abiotic factors have on these and other marine metabolites and their measurements in the ocean.

Photodegradation experiments
Thirty-minute experiment: We added 8 pM of Me-and Ado-, and 5 pM of CN-B 12 and OH-B 12 into 15 mL of seawater, collected at HL05, individually in ultraviolet (UV)-penetrable quartz vials (QP059, Cuvet.Co).Duplicate samples (15 mL) were taken directly after the addition of cobalamins at time 0.Then, duplicate quartz vials were exposed to direct, natural sunlight then 15 mL sample was transferred into clean amber vials in the dark at À20 C after 5, 10, 15, and 30 min on 07 June 2022 (Ado-, Me-, and CN-B 12 ) (35.1 C; 1450 μmol s À1 m À2 ) and 21 July 2022 (OH-B 12 ) (26.6 C; 1400 μmol s À1 m À2 ).Four-day experiment: We added 4 pM OH-B 12 in 20 mL seawater, collected at HL12, in quartz vials, took duplicate T 0 samples then exposed to direct sunlight in May 2022.We transferred duplicate vials daily at noon into clean amber vials then stored in the dark at À20 C until processing.Temperature (Logger, RC-5, Elitech) and light (Digital Light Meter, LUX29TK, TekcoPlus) were monitored (Supporting Information Table S2).For the 4-d experiment, mean day length was 11 h, mean air temperature was 23.1 C, and average daytime light $ 950 μmol s À1 m À2 with maximum 1400 μmol s À1 m À2 .We collected dark controls that were supplemented with the same concentrations of cobalamin forms, wrapped in opaque black bags, exposed to similar temperatures, and sampled identically.We also processed control seawater samples in an identical manner to determine endogenous concentrations of cobalamins (Supporting Information Table S2).
We performed all extractions in a dark, windowless room with only a red headlamp as a light source.Samples were extracted on 100 mg HyperSep C18 solid-phase extraction (SPE) cartridges (ThermoScientific, 60108-302) using a vacuum manifold.SPE cartridges were preconditioned with 2 Â 0.5 mL MeOH then 2 Â 0.5 mL deionized water and kept wet through the entire protocol.Samples were extracted at $ 1 mL min À1 , washed with 2 Â 0.5 mL milli-Q water then eluted with 2 Â 0.85 mL MeOH into clean microtubes.Eluent was dried for $ 2 h under vacuum (Vacufuge, Eppendorf), in the dark, then stored at À80 C prior to analysis when it was resuspended with 100 μL of H 2 O containing 0.1% formic acid.

Degradation product identification experiment
To investigate OH-B 12 degradation products, we prepared 1 μM of OH-B 12 in HPLC grade water and exposed it to ambient sunlight and temperature for 8 d (EXP 1) and 4 d (EXP 2).One sample from EXP 2 on day 4 was exposed to two 5-min intervals of 254 nm ultraviolet-C (UVc) light from a 30-W lamp (EXP 2, day 4 + UV).We diluted samples 200-fold before LC-MS analysis as outlined below and in Supplemental Methods.

Mass spectrometry
Cobalamin photodegradation in seawater was quantified using a Dionex Ultimate-3000 LC system coupled to an electrospray ionization source of a TSQ Quantiva triple-stage quadrupole mass spectrometer (ThermoFisher) with transition list reported in Supporting Information Table S4.Photodegradation products were investigated on an Agilent 1290 Infinity II LC coupled to a Q Exactive HF Orbitrap mass spectrometer (ThermoFisher) with a heated electrospray ionization probe using both data-dependent acquisition and full scan modes.Complete details of LC-MS analyses are provided in Supplemental Methods.

Cobalamin quantification
We combined equal portions of each sample to obtain quality control (QC) pools for the 30-min and 4-d experiments.Cobalamins were quantified using standard addition with calibration curves prepared in the QC pools using the LC-MS approach described in the Supplemental Methods.Duplicate injections were performed with 0, 2.5, 5, and 25 fmol on C18 column for all cobalamin forms (Supporting Information Fig. S7) and R 2 for slopes were all >97%.SPE percent recovery for 100 mg HyperSep C18 cartridge (n = 2) was determined by spiking 15 mL seawater samples with each cobalamin form equivalent to 80 fmol on C18 column before or after preconcentration on SPE.All reported concentrations and limits of detection are corrected for percent recoveries.Raw files generated with Xcalibur software (ThermoFisher) were uploaded into Skyline Daily (University of Washington) and the transitions with the best signal-to-noise and lowest interference were selected for quantification (Supporting Information Table S4).Limits of blanks (LOB) and Limits of detection (LOD) were determined according to Armbruster and Pry (2008) (Supporting Information Table S3).

Photodegradation product identification
We identified potential photodegradation products of OH-B 12 in pure water using three approaches.First, we selected peaks of interest based on their increasing peak area in samples during sunlight exposure by performing a "background subtract" of the first samples (T 0 ) from T 4 and T 8 samples.Second, we assembled reports of cobalamin degradation products from the literature and searched for their accurate masses (AE5 ppm) (Supporting Information Table S5).Third, we searched product ion spectra for masses that are known cobalamin fragments (m/z 147.0922 and m/z 359.1005 product ions, DMB and DMB sugar-phosphate, respectively).All candidate products were absent in blank samples.
We used exact mass to determine chemical formula of candidate degradation products by constraining upper limits with ring-double bond equivalent (< 26), elemental composition (C 62 H 90 CoN 13 O 15 P), and MS run-specific mass error ($ 4.0 ppm) of the OH-B 12 standard.We present level 2 (diagnostic) probable chemical structures (Fig. 3) if they were either previously reported in literature or we were able to obtain a conclusive product ion spectrum (Schymanski et al. 2014).

Statistical analysis
Two outliers (Ado-B 12 : 15 min, CN-B 12 : 30 min) were removed from the photodegradation experiment analyses during data analysis due to abnormally high B 12 concentrations (Grubbs test, p-value < 0.05), likely due to contamination during extraction (Supporting Information Fig. S1).Differences between T 0 and dark controls at 30 min and 4 d (Supporting Information Table S4) as well as OH-B 12 at T 0 and T 30 (Fig. 2D) were normally distributed and determined insignificant using a student t-test (Supporting Information Table S2).

Results
During the photodegradation experiment, Me-B 12 (Fig. 2A) and Ado-B 12 (Fig. 2B) were not detected after 5 min in sunlight.Approximately, 35% of CN-B 12 (Fig. 2C) remained after 30 min of sunlight exposure.Approximately, 63%, 57%, and 96% of Me-B 12 , Ado-B 12 , and CN-B 12 photoconverted into OH-B 12 by the end of the 30-min experiment (after corrections for endogenous OH-B 12 concentrations present at the beginning of the time course, Supporting Information Table S2).No significant degradation of OH-B 12 was detected over 30 min (Fig. 2D, p-value = 0.6449) but only approximately 18% of OH-B 12 added remained after 4 d of light exposure (Fig. 2E).No significant degradation of Me-B 12 , Ado-B 12 , CN-B 12 (Supporting Information Table S2), and OH-B 12 (Fig. 2E [white squares]; Supporting Information Table S2) was detected in controls exposed to the same temperature but in the dark.Endogenous OH-B 12 concentrations ranged from < LOD to 2.7 pM in seawater samples (Supporting Information Table S2).Endogenous concentrations of Me-B 12 , Ado-B 12 , and CN-B 12 were < LOD in samples prior to spiking (Supporting Information Table S2).
Four potential photodegradation products of OH-B 12 were characterized because of their increasing peak area over time in pure water when exposed to sunlight.Two products were identified based on accurate mass because they were previously described as products of cobalamin acid hydrolysis: a part of the corrin ring, 3,3-dimethyl-2,5-dioxopyrrolidine-4-propionamide (C 9 H 15 O 3 N 2 + , [M + H] + m/z 199.1077) (Fig. 3B) (Kuehl et al. 1955) , S3).Two larger degradation products were identified in this study that, to our knowledge, have not been reported.A product with an [M + H] + at m/z 640.2378 has a chemical formula of C 27 H 39 O 11 N 5 P + was identified through its similar fragmentation as OH-B 12 .Product ion spectra confirmed the presence of both m/z 147.0922 and m/z 359.1005 product ions (DMB and DMB sugar-phosphate, respectively) (Fig. 3D; Supporting Information Fig. S4).The largest candidate degradation product detected had an [M + H] + at m/z 832.3641 and a chemical formula of C 38 H 55 O 12 N 7 P + .Product ion spectra confirmed the presence of both m/z 147.0922 and m/z 359.1005 product ions (DMB and DMB sugarphosphate, respectively) (Fig. 3E).

Timescales of cobalamin forms photodegradation in seawater
Me-, Ado-, and CN-B 12 photodegradation timescales in seawater (Fig. 2A-C) are similar to previous measurements in blood and plasma despite chemical differences in matrices (Ahmad et al. 1992;Juzeniene and Nizauskaite 2013;Vaid et al. 2018).However, OH-B 12 degradation measured here was notably faster compared to the sole degradation study in seawater, performed by Dr. Carlucci in 1969 (Carlucci andSilbernagel 1969).Here, only $ 18% of OH-B 12 was present after 4 d (Fig. 2E) while approximately 50% of cobalamin activity, assessed by bioassay, remained after 7 d in the earlier study by Carlucci and Silbernagel (1969).These differences may be attributable to the different analytical approaches.While our targeted mass spectrometry method quantifies only a single specific molecule, bioassays likely quantify the presence of unknown cobalamin degradation products that are still of utility to cells for meeting cobalamin demand.
Additional differences in these observed timescales of degradation could also arise from environmental factors such as light intensity (Ahmad and Hussain 1993;Kumar and Kozlowski 2012).However, the 1969 study by Carlucci et al. (Carlucci and Silbernagel 1969) had approximately double the maximum light  S2.LOD and percent recoveries are shown in Supporting Information Table S3.
exposure, 1800 μmol s À1 m À2 , compared to our 4-d study, which had an average of 950 μmol s À1 m À2 (Carlucci and Silbernagel 1969).The chemical composition of seawater may also influence cobalamin stability directly and also alter the penetration of sunlight into a sample.In the 1969 study, seawater was collected off the Scripps Institution of Oceanography pier while seawater used here was collected from the open ocean (Sta.HL12).The penetration of UV and visible light in filtered seawater depends on light interactions with dissolved organic matter (Tedetti and Sempéré 2006), which was most likely higher in the pier-collected, coastal water.In addition, OH-B 12 may degrade quicker in the presence of reactive oxygen or nitrogen species common in the ocean (Manzanares and Hardy 2010) which could have been variable between these locations.Riboflavin, which has been measured at up to 160 pM in the ocean, also enhances degradation of cobalamin (Ahmad 2012;Heal et al. 2014).Just as these factors could explain differences in degradation between experiments, they also likely play a role in the field.To acknowledge this variability, rather than reporting specific degradation rates, we instead provide general timescales of photodegradation (Fig. 4).Future work in identifying the dominant factors that influence OH-B 12 photodegradation rates is an important next step to understand and predict spatiotemporal patterns of cobalamin availability.The rapid photodegradation of Ado-, Me-, and CN-B 12 into OH-B 12 seeds doubt that any form other than OH-B 12 would dominate the pool of dissolved cobalamin in euphotic zone of the ocean.This, however, is at odds with literature reports where 58% of studies report  as the dominant form and 43% report Me-B 12 (3/7) as dominant (Fig. 1B).Adoand Me-B 12 could be detectable during periods of extremely active biosynthetic production, high biomass turnover, or mass senescence events, coupled with complete lightprotection during sample collection and processing.This emphasizes the importance of careful sampling procedures coupled to characterization of ecosystem productivity and biomass to fully interpret cobalamin-related compound measurements.However, the studies that report the dominance of these labile forms do not describe intensive measures to protect from light beyond using amber bottles to store samples.This suggests that these reports of Ado-or Me-B 12 dominance could be artifacts arising from difficulties involved in LC-MS analysis or quantification methods.For example, these molecules are often quantified using much more photo-stable compounds (such as B 2 ) as internal reference standards, leveraging response factor calculations for quantitative reporting (Suffridge et al. 2017(Suffridge et al. , 2018)).This photostability mismatch has the potential to distort measurements of the photo-labile (Ado-and Me-B 12 ) and photo-stable cobalamins.In addition, calculating concentrations from response factors are difficult to replicate without exceptionally detailed experimental procedure reporting and data transparency.Quantification based on calibration curves performed in water instead of representative, pooled samples are common (Suffridge et al. 2017;Möller et al. 2022) and could significantly alter measurements due to intense matrix effects (Heal et al. 2014;Boysen et al. 2018).Third, Me-and Ado-B 12 standards are also subject to this rapid photodegradation and if great care is not taken to protect them from light during calibration curve construction, this degradation could lead to an overestimate of in situ concentrations.In addition, enhanced photodegradation is more likely if SPE extractions are performed at sea, a less controlled environment in general, compared to a land-based laboratory.The higher stability of CN-B 12 could lead to its detectable presence in the ocean, particularly at increased depths, regions with high anthropogenic input (coastal) or, potentially, areas with high cyanide concentrations where spontaneous production of CN-B 12 from OH-B 12 could occur (Paul and Brady 2017).Future efforts toward intercalibration and standardization of cobalamin and related compound measurements in the water column should be prioritized.
Finally, a recent publication suggested that the dominant form of cobalamin in the North Sea was a putative OH-B 12 isoform that had similar fragmentation to OH-B 12 but eluted later (Möller et al. 2022).We have not detected this putative isoform here or in any marine water column samples to date, despite our method's suitability for detecting this putative isoform's reported molecular ion in the proposed elution order (Supplemental Methods; Fig. S6).

Cobalamin degradation products
It is estimated that up to 17% of bacteria can salvage cobalamin intermediates for repair and use to satisfy cobalamin demand, based on their expression of partial genetic pathways in cobalamin biosynthesis (Shelton et al. 2019).However, cobalamin precursor or degradation products have not been identified in the environment and our current understanding of their potential microbial use is largely limited to documented corrinoid salvaging and pseudocobalamin remodeling (Yi et al. 2012;Helliwell et al. 2016;Ma et al. 2020).DMB is present in three of the probable photodegradation products measured here, including alpha-ribazole (Fig. 3C,D; Supporting Information Fig. S5).Alpha-ribazole was identified in 1950 as an acid-hydrolysis product from cobalamin and recently has been hypothesized to be produced by select marine bacteria and influence marine prokaryotic community composition and growth (Johnson et al. 2016;Wienhausen et al. 2017Wienhausen et al. , 2022)).In B 12 synthesis pathways, alpha-ribozole is derived from DMB (lower ligand, Fig. 1A) and could be incorporated into B 12 at its lower ligand (Rodionov et al. 2003;Lu et al. 2020).This could be important to organisms that are able to remodel cyanobacterial produced pseudocobalamin (psB 12 ) to B 12 by switching the lower ligands (Helliwell et al. 2016).Here, we provide evidence that alpha-ribozole is a photodegradation product of OH-B 12 .Measuring the prevalence of these degradation products, and others, in natural environments and assessing their bioavailability to cultures of microbes is a key next step.
None of the photodegradation products we detected contained a cobalt ion; however, our analytical approaches were in no way exhaustive: the ability to detect unknown compounds, including those containing cobalt, is strongly biased by our selection of chromatography gradients, columns, and ionization modes.A recent study, Wong et al. (2022).speculated that although cobalamin itself was not a significant contributor to the pool of cobalt-binding ligands, cobalamin degradation products could be warranting further investigation.

Conclusion
Cobalamin is an important micronutrient that influences marine microbial communities' activity and interactions, but its photolability has largely been overlooked in marine studies.The photodegradation products identified here advance efforts to understand microbial use of cobalamin intermediates.Furthermore, cobalamin form-specific photodegradation timescales (1) inform our understanding of the dominant forms of cobalamin we expect in the sunlit ocean and (2) highlight the importance of avoiding light exposure during sampling, extraction, and quantification.Adopting light protective measures will yield more accurate and reliable cobalamin measurements from which to examine spatiotemporal patterns in cobalamin dynamics, important considerations for measurements of marine metabolites more broadly.

Fig. 1 .
Fig. 1.A survey of dissolved cobalamin forms and concentrations documented in the ocean.(A) General classification of four cobalamin forms.(B) Global distribution of published dissolved B 12 measurements in seawater in chronological order.Bioassay-based measurements are represented with black circles.Colored circles represent the cobalamin forms detected in the surface ocean in order of increasing dominance from left to right.(C) Average total dissolved cobalamin concentration (pM) on log scale in surface ocean samples.Numbers refer to published studies as in (B).Study details and cobalamin concentrations can be found in Supporting Information TableS1.

Fig. 2 .
Fig. 2. Photodegradation of (A) Me-B 12 , (B) Ado-B 12 , (C) CN-B 12 , and (D, E) OH-B 12 spiked into seawater in natural sunlight over 30 min (A-D) and 4 d (C).Photoconversion of (A) Me-B 12 , (B) Ado-B 12 , and (C) CN-B 12 into OH-B 12 (white circles).(E) OH-B 12 in comparable temperature but in the dark over 4 d (white squares).Gray dashed line represents pM LOD, adjusted for percent recovery, for R-B 12 forms (A-C) and OH-B 12 (D, E), points falling below LOD were set to equal zero.Points represent pM value for each replicate vial (n = 2, for most).Dark controls, initial values and environmental factors are shown in Supporting Information TableS2.LOD and percent recoveries are shown in Supporting Information TableS3.

Fig. 4 .
Proposed photodegradation time-scale of cobalamin forms in seawater.

Fig. 3 .
Fig. 3. (A) LC-HRMS characteristics of OH-B 12 degradation products and changes in their relative abundance during two degradation experiments over