4‐Siderophore property statistics
The content of this page is the supplementary materials of our paper. If you find it useful, please cite it.
Figure S1. The statistics of 649 unique siderophores in SIDERITE
A. Distribution of the siderophore producers by their kingdoms.
B. Distribution of the siderophore biosynthetic pathways.
C. Distribution of the functional group type combinations. For clarity, only the top ten combinations are shown, and the others are merged into "Others".
D. Distribution of the common functional group of siderophores. One siderophore could contribute to more than one functional group type if it contains many types of functional groups.
E. Distribution of denticity numbers.
F. Distribution of the molecular weight.
The 649 unique siderophores in our database can be classified by their producer sources (Figure S1A).
At the kingdom level, the majority of 649 siderophores in the database are produced by bacteria (85.90%), followed by fungi (12.40%), plants (1.54%), and animals (0.15%).
While most siderophores are specific to one producer source at the kingdom level, there is one exception.
Triacetylfusarinine can be produced by both bacteria (Paenibacillus triticisoli) and fungi (Penicillium sp.).
At the phyla level, 872 producers of 649 siderophores are spread across five major bacterial phyla (229 in Actinobacteria, 7 in Cyanobacteria, 33 in Firmicutes, 412 in Proteobacteria, and 8 in Bacteroidetes) and three major fungal phyla (124 in Ascomycota, 38 in Basidiomycota and, 8 in Mucoromycota, Table S1).
Siderophores can also be classified by their biosynthetic pathways (Figure S1B).
Previous studies have suspected that the NPRS pathway is more dominant for siderophore synthesis[8], yet no statistically concrete ratio between NRPS and NIS-derived siderophore has been established.
In SIDERITE, we found that related pathways can be classified into NRPS (65.18%), NIS (21.73%), Hybrid NRPS/PKS (10.02%), Hybrid NRPS/NIS (2.77%), and PKS (0.31%).
Consistent with previous studies, NRPS was indeed the most abundant biosynthetic pathway of siderophores, followed by NIS. PKS siderophores are rare, with only two cases (proferrorosamine A and tetracycline).
In different kingdoms, the composition of the biosynthetic pathway differs (Table S3).
In plants and animals, all siderophores are synthesized by NIS. In fungi, almost all siderophores are synthesized by NRPS (90.12%, 73/81).
In bacteria, the diversity of the siderophore biosynthetic pathways is higher, with 62.90% (351/558) NRPS, 22.58% NIS (126/558), 11.29% hybrid NRPS/PKS (63/558), and 2.87% hybrid NRPS/NIS (16/558).
Digitization also enables us to easily access the statistical properties of siderophores, such as functional group distribution.
Siderophores chelate iron by several common functional groups (coordinating groups) [8], and a single siderophore may use multiple types of functional groups.
In SIDERITE, the top five combinations of functional groups are present in 64.25% of the siderophores (Figure S1C).
These top five combinations are Hydroxamate (28.35%), Alpha-Hydroxycarboxylate+Hydroxamate (12.02%), Catecholate+Hydroxamate (8.78%), Hydroxamate+Hydroxyphenyloxazoline (8.63%) and Catecholate (6.47%).
Hydroxamate and Catecholate are the most common functional groups found in 69.49% and 29.74% of siderophores, respectively (Figure S1D).
Most of siderophores have three bidentate groups which forming octahedral geometry with iron in coordinating number six [8].
As expected, most siderophores have a denticity number of six (Figure S1E).
However, there are exceptions, such as pacifibactin, desferrioxamine T1, malleobactin D and pyoverdine 7.7, which contain four bidentate groups.
It’s reported that pacifibactin coordinates iron in a 1:1 ratio, and the function of the extra coordinating group is unknown [9].
The molecular weight (MW) is another important property of siderophores, as it influences the diffusion of siderophores [10].
Siderophores are typically small molecules, and the MW of siderophores in our database ranges from 138.12 Da (salicylic acid) to 1766.86 Da (pyoverdine IB3).
Most siderophores (90.60%, 588/649) have a middle MW ranging from 300~1100 Da (Figure S1F).
Of note, half of the smallest 12 siderophores (MW < 200 Da) are monomers of other siderophores, such as salicylic acid, 2,3-dihydroxybenzoic acid, and citrate.
Most of the large siderophores (MW > 1200 Da) are pyoverdines (87.76%, 43/49), as pyoverdines are frequently composed of more than ten amino acids.
Further, we check whether MW distribution is related to biosynthetic types.
We found NRPS siderophores exhibit a wide range of MW, from 206.20 Da (Spoxazomicin D) to 1766.86 Da (pyoverdine IB3).
NRPS siderophores are generally heavier (mean: 835.68 Da; median: 816.00 Da) than NIS siderophores (mean: 506.93 Da; median: 516.63 Da).
**Figure S17 The predicted properties of 649 siderophores.**
(A) The predicted aqueous solubility. The unit is log10 of solubility in the water.
(B) The predicted diffusion coefficient. The temperature is set to 298.15K and the solvent is water.
Aqueous solubility and the diffusion coefficient also are important ecological properties of siderophores.
However, most siderophores have not been measured experimentally for these two properties.
Therefore, we employed SolTranNet [2] and Stokes–Einstein Gierer-Wirtz Estimation (SEGWE) [3] to predict these properties using SMILES representations (Figure S17).
Out of the total 649 siderophores, 635 were predicted to be water-soluble (predicted logS>-6).
Among them, 551 siderophores exhibited a “good” level of aqueous solubility (predicted logS>-4).
Notably, 14 siderophores exhibit poor aqueous solubility (predicted logS≤-6) due to the presence of long-chain fatty acids, suggesting them are probably insoluble in water.
The minimum and maximum predicted diffusion coefficients are 2.66×10-10 m2/s and 7.7×10-10 m2/s respectively.
The majority of siderophores (611 out of 649) exhibit predicted diffusion coefficients within the range of 2.66 to 5.50×10-10 m2/s, consistent with previous research [11].
Figure S18 The distribution of C/N and C/O ratios in the different biosynthetic types and clusters.
(A) The distribution of C/N ratio between different biosynthetic types of siderophores.
(B) The distribution of C/O ratio between different biosynthetic types of siderophores.
(C) The distribution of C/N ratio between different clusters of siderophores. (D) The distribution of C/O ratio between different clusters of siderophores.
Molecule weight (MW), carbon-nitrogen ratio (C/N), and carbon-oxygen ratio (C/O) of siderophores pertain to their biosynthetic cost in various aspects: higher MW can indicate more building blocks; while producing products with lower C/N or C/O could influence growth under nitrogen-limited or hypoxic environments [12−14].
Therefore, we investigated these properties of siderophores in different clusters (Figure S18).
Siderophores in cluster 4 have significantly higher MW than average, as most cluster members are pyoverdines.
Regarding C/N and C/O ratios, different clusters also exhibit different properties.
For example, siderophores in cluster 7 and cluster 10 have significantly higher-than-average C/N because they only contain 1~2 nitrogen atoms, and siderophores in cluster 4 have a significantly lower-than-average C/N (Figure S19).
Figure S19 The distribution of nitrogen atom and oxygen atom numbers in the different biosynthetic types and clusters.
(A) The distribution of nitrogen atom number between different biosynthetic types of siderophores.
(B) The distribution of oxygen atom number between different biosynthetic types of siderophores.
(C) The distribution of nitrogen atom number between different clusters of siderophores.
(D) The distribution of oxygen atom number between different clusters of siderophores.
All groups encompass siderophores derived from unique biosynthetic types, except groups 1.5 and 1.10, where siderophores are synthesized through both hybrid NRPS/PKS and NRPS pathways.
The siderophores in these two groups have two biosynthetic types due to the optional modification from PKS.
Also, siderophores within the same group are typically produced by species from the same kingdom.
However, groups 2.1, 2.9, 3.7, and 9.1 contain producers from both bacteria and fungi, and group (11.1) is produced by bacteria and animals.