1
Supplementary text S1:
2
EvoMining of Streptomyces sviceus draft genome reveals an Enolase enzyme family member
3
recruited into a new phosphonate BGC
4
Enolase is a glycolytic enzyme that catalyzes the dehydration of 2-Phosphoglycerate (2-PGA) to
5
produce phosphoenolpyruvate (PEP) in a Mg++-dependent reaction. The enolase phylogeny (Tree
6
S2) has two main clades; a major clade that includes orthologs associated with central metabolism
7
from representatives of most species in the genome database (red braches, Figure S1A). As
8
expected, the general topology of this clade reflects that of the guide species tree (Tree S1). A
9
divergent clade (cyan, blue and green branches, Figure S1A) includes a homolog from Streptomyces
10
viridochromogenes that has previously been identified found in the BGC for the phosphinothricin
11
tripeptide (PTT) (1). This clade also includes a homolog from Streptomyces sviceus (GI 297146550;
12
eno2-SSV) that has not been linked to NP biosynthesis.
13
The S. viridochromogenes PTT enolase or carboxyphosphoenolpyruvate synthase (CPS GI:
14
302549806) shares 33% sequence identity with its glycolytic counterpart, i.e. GI 302551949. A
15
detailed sequence analysis showed only few changes in the active site residues (Figure S2A). To
16
identify the tridimensional position of these changes, a structural model of eno2-SSV was obtained
17
and compared with the crystal structure of the yeast enolase (PDB: 2ONE), which has been
18
thoroughly characterized (2). This sequence and structural analysis revealed that the mutation
19
E211S (numbering of yeast enolase) affects the active site of CPS. To analyze the effect of this
20
mutation, the CPS substrate, 2-Phosphonoformylglycerate was modeled in the active site of both
21
structures. This analysis showed that the ancestral glutamine residue would not allow the
22
accommodation of the substrate (Figure 2SB). Therefore, this particular mutation seems key to
1
23
substrate specificity in CPS. Overall, this analysis suggests that other members of the divergent
24
clade are related to a new enzyme function, likely involved in NP biosynthesis.
25
The draft genome sequence of S. sviceus has been deposited as a single scaffold with 551 gaps
26
(GenBank accession: CM000951.1 and BioProject PRJNA59513). Six gaps were located in the
27
region of interest, including one at the 5’ end of the enolase homolog, leading to a partial sequence.
28
Remarkably, neither PKSs nor NRPSs could be found in the gene neighborhood of the CPS gene,
29
although an incomplete CDS for a mutase resembling those related to phosphonates (1) could be
30
detected. On the basis of the phylogenetic analysis we expected that the divergent clade includes
31
enolases that are part of a BGC. To confirm this, the six gaps in the region were closed by iterative
32
PCR amplification (Supplementary Text 1 associated Table 1) and sequencing, followed by manual
33
annotation of the region. The annotation and functional predictions confirmed the presence of a
34
BGC putatively encoding a pathway that shares common steps with PTT biosynthesis, including
35
those related to the formation of phosphinopyruvate from phosphoenolpyruvate (1) (Supplementary
36
Figure 2C). Moreover, the complete sequence allowed for the identification of the mutase-
37
decarboxylase pair of enzymes present in most Streptomyces phosphonate biosynthetic systems
38
(Figure S1B). Overall, this functional annotation suggests that the product of this BGC is a
39
previously uncharacterized phosphonate natural product.
40
41
42
43
44
2
45
Supplementary text 1 associated table 1.
46
Annotation of a new phosphonate BGC in S. sviceus
47
Gene
Locus tag
Predicted function
Length
(Amino acids)
Closest homolog
ID*
1
SSEG_06268
LysR family transcriptional
regulator
306
LysR family transcriptional regulator, Nostoc
punctiforme PCC 73102
36%
2
SSEG_09941
ABC transporter ATPase
subunit
326
ABC transporter, Catenulispora acidiphila DSM
44928
63%
3
SSEG_09940
ABC-type multidrug transporter
permease
261
ABC transporter, Catenulispora acidiphila DSM
44928
63%
4
SSEG_06265
Phosphonate dehydrogenase
372
D-isomer specific 2-hydroxyacid dehydrogenase,
Cyanothece sp. ATCC 51142
45%
5
SSEG_09939
Phosphopantetheinyl
transferase
223
4'-phosphopantetheinyl transferase, Nocardia
asteroides
39%
6
SSEG_09938
Phosphopantetheine-binding
protein
106
Hypothetical protein, Actinokineospora
enzanensis
46%
7
SSEG_06262
Phosphonate-acyltransferase
556
Hypothetical protein, Salinispora pacifica
50%
8
SSEG_06261
Manganese transporter MntH
439
mn2+/fe2+ transporter, nramp family,
Micromonospora sp. L5 YP_004081943.1
52%
9
SSEG_09937
Phosphoenolpyruvate
phosphomutase
309
Phosphoenolpyruvate phosphomutase,
Saccharopolyspora spinosa
61%
10
SSEG_09936
Rieske (2Fe-2S) iron-sulfur
domain protein
126
Rieske (2Fe-2S) iron-sulfur domain-containing
protein, Pseudonocardia dioxanivorans CB1190
46%
11
SSEG_09935
Metallo-dependent
amidohydrolase
361
Hypothetical protein, Paenibacillus daejeonensis
50%
12
SSEG_09934
Short chain
dehydrogenase/reductase family
284
Alcohol dehydrogenase, Nocardiopsis
halotolerans
44%
13
SSEG_09933
Glutamate-1-semialdehyde
aminotransferase
468
Glutamate-1-semialdehyde aminotransferase,
Pseudomonas mendocina NK-01
36%
14
SSEG_09932
Aminolevulinate-coenzyme A
ligase
412
8-amino-7-oxononanoate synthase, Pontibacter
sp. BAB1700
58%
15
Within a gap
Putative alcohol dehydrogenase
382
Alcohol dehydrogenase, Streptomyces rimosus
41%
16
SEG_10418
Hydroxyethylphosphonate
dioxygenase
439
2-hydroxyethylphosphonate dioxygenase phpD,
Streptomyces viridochromogenes
46%
17
SSEG_10417
3-phosphoglycerate
dehydrogenase
338
D-3-phosphoglycerate dehydrogenase, Frankia
alni ACN14a
62%
18
Within a gap
Phosphonopyruvate
decarboxylase
383
phosphonopyruvate decarboxylase, Nocardia
brasiliensis ATCC 700358
63%
19
Within a gap
Nicotinamide mononucleotide
adenylyltransferase
184
Nicotinamide mononucleotide adenylyltransferase
phpF, Streptomyces viridochromogenes
74%
20
SEG_10416
2,3-bisphosphoglycerateindependent phosphoglycerate
mutase
427
PhpG, Streptomyces viridochromogenes
61%
21
SSEG_10415
Enolase
421
Carboxyphosphoenolpyruvate synthase,
Streptomyces viridochromogenes
50%
22
SSEG_10414
Carboxyphosphonoenolpyruvate mutase
286
Carboxyphosphonoenolpyruvate mutase,
Streptomyces hygroscopicus
80%
23
SSEG_10413
Aldehyde dehydrogenase
462
Hypothetical protein, Amycolatopsis nigrescens
48%
24
SSEG_08119
Beta-lactamase domaincontaining protein
255
Aldehyde dehydrogenase PhpJ, Streptomyces
viridochromogenes
69%
*Percentage of amino acid sequence identity based in the BlastP alignment
3
48
Supplementary text 1 associated methods.
49
Streptomyces sviceus BGC gap closure. The gaps and misassembles found in the region between
50
8 and 24 kbp downstream and upstream of the PTT enolase (ZP_06914376.1) in the S. sviceus draft
51
genome sequence, which was obtained from GenBank (GI: NZ_ABJJ00000000), were closed by
52
PCR amplification and product sequencing (Supplementary text 1 associated table 2); for gap 3,
53
which was too long for a single PCR, 3 iterative rounds of sequencing and primer synthesis were
54
required until the gap was closed.
55
Molecular modeling of the recruited enolases (PhpH). The molecular model of PhpH was
56
constructed with Modeller (3) using as template the crystal structure of the dimeric yeast enolase in
57
complex with magnesium, 2-phosphoglycerate (2-PGA) and phosphoenolpyruvate (PEP) obtained
58
from the Protein Data Bank (PDB : 2ONE) (2). This enolase shares 33% identity with the Carboxy-
59
phosphoenolpyruvate synthase (phpH or PTT enolase) from S. viridochromogenes (1). A model of
60
the product of the PTT enolase, carboxy-phosphoenol pyruvate was built with VegaZZ (4), and
61
located in an analog position with respect to PEP in the active site of the PTT enolase, by means of
62
using superimpositions of the model and template in Pymol (The PyMOL Molecular Graphics
63
System, Version 0.99 Schrödinger, LLC; http://www.pymol.org/).
64
Supplementary text 1 associated table 2. Primers used for gap closing
Fragment
1
2
3
4
5
6
Total sequenced
Forward primer
Reverse primer
bases
148
F-TGCCGCCCAGTTCGAGCAGA R-ATCCGAACGCACACCGCTG
566
F-CCAGCGTTCTGGCCAGGGCT R-CACGATCGCGACCGACGACT
FA-AAGGCGCCCTGCTTGATGAA RA-CAAACTCCAGGCCTTCTACG
FFB-GAAGTTGATGCGGAACGCCA RB-GCCGAGAACATCCTGCACGTG
2726
FC-GCTGATGGGTTTGTCGTCGC RC-GGTGGCGTGATGGTCACAGC
RD-CGTGTGCACCACCGGCAAGTC
538
F-ATTCCGGTTGTTGGCGTGCC; R-TAGTTGTTGATGCTCCACAC
484
F-GTCGTCGAAGTCATGGGCGT; R-CATGGTCTTCGACACCCTGG
535
F-GAGTGGTCGGCATGGGCCGG; R-GTGACCTCGTGATCCGGGAC
65
4
66
67
68
69
70
71
72
73
74
75
76
77
78
79
Supplementary text 1 associated references:
1. Blodgett JA et al. (2007) Unusual transformations in the biosynthesis of the antibiotic
phosphinothricin tripeptide. Nat Chem Biol 3:480–5
2. Zhang E, Brewer JM, Minor W, Carreira LA, Lebioda L (1997) Mechanism of enolase: the crystal
structure of asymmetric dimer enolase-2-phospho-D-glycerate/enolase-phosphoenolpyruvate at 2.0
A resolution. Biochemistry 36:12526–34.
3. Sali A, Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol
Biol 234:779–815.
4. Pedretti A, Villa L, Vistoli G (2004) VEGA--an open platform to develop chemo-bio-informatics
applications, using plug-in architecture and script programming. J Comput Aided Mol Des 18:167–
73.
80
5
81
82
Figure S1. A. Phylogenetic reconstruction of the actinobacterial enolases (Tree S2). Black branches
83
include homologs associated with glycolysis while green branches were linked to NP BGCs, a
84
homolog from S. sviceus, highlighted in red implicates the loci shown in B in phosphonate
85
biosynthesis. B. The gene cluster (top) that encodes a novel biosynthetic pathway for a cryptic
86
phosphonate NP identified using EvoMining on the genome of S. sviceus. The gene cluster
87
organization is compared with the PTT gene cluster of S. viridochromogenes. At the bottom the
88
common biosynthetic steps between the PTT and PSV pathways are shown
89
6
90
91
Figure S2. Structure-function analysis of enolases and carboxyphosphoenolpyruvate
92
synthases (CPS). A. Sequence alignment of enolases from various organisms and CPS, the amino
93
acid numbers are relative to the yeast enolase. The catalytic residues are indicated at the top and
94
central homologs are shown in white background, and recruited homologs in green as in the
95
phylogenetic reconstruction in supplementary figure 1A. B. Comparison of the yeast enolase crystal
96
structure bound with its product phosphoenolpyruvate (PEP) and a structural model of the CPS from
97
S. viridochromogenes and its substrate carboxy-phosphoenolpyruvate (CPEP), K345 the conserved
98
catalytic base, and the mutations in the catalytic acid E211S and the catalytic water molecule holder
99
E168Q are indicated and shown in sticks. C. Reactions catalysed by the glycolytic enolases and the
100
CPSs, colour code is the same as in A and supplementary figure 1A.
101
7
102
103
104
Figure S3. Distribution of EvoMining hits by BGC class as annotated by AntiSMASH. The most
105
abundant known classes of BGCs are NRPSs (23%) and PKSs (PKS1, PKS2, PKS3 and TransPKS;
106
18 % in total). EvoMining predictions and EvoMining hits detected by ClusterFinder are altogether
107
the most abundant class (30 %) and may represent several classes of unprecedented BGCs
108
8
109
110
Supplementary figure S4. A. HPLC analysis (Vydac C-18 column) of extracts of a leupA deficient
111
mutant in S. roseus ATCC31245 in comparison with wild type S. roseus and a leupeptin authentic
112
standard, leupeptin can be detected (see figure S5 for MS analyses) in wild type S. roseus while the
113
leupA mutant cannot longer produce leupeptin. B. HPLC analysis (Restek C-18 column) of extracts
114
from E. coli DH10B carrying the 8_10B and 9_18N clones with the leup locus in comparison with
115
a leupeptin authentic standard. Both strains produced leupeptin (see figure S5 for MS analyses).
9
116
117
118
119
120
Figure S5. A. MS analysis of peaks with retention times equivalent to the leupeptin standard (See
121
figure S4) confirming heterologous production of leupeptin using genomic clones containing the
122
leup genes. B. MS2 analysis of genomic clones containing the leup genes, the fragmentation patterns
123
of the m/z =427.3 from the extracts are identical to those of the m/z=427.3 from the leupeptin
124
standard. Similar patterns were obtained from the extracts of wild type S. roseus ATCC 31245.
10
125
126
Figure S6. Postulated pathway for arseno-organic NP biosynthesis in S. coelicolor and S.
127
lividans. The reactions proposed for SLI_1096, SLI_1097 and SLI_1091 are responsible for the
128
biosynthesis of the As-C bond at the early stages of the biosynthetic pathway. The biosynthetic logic
129
proposed for SLI_1088-9 is related to the synthesis of an acyl chain that is proposed to be linked to
130
the As-C containing intermediary by other enzymes in the BGC. At the left, structural predictions
131
of potential products for the pathway based on high resolution MS data are shown. This pathway
132
and further studies on the water-soluble As-species present in the samples (data not shown) suggest
133
a non-methylated As-moiety as shown in the last structure, which has not been described in literature
134
yet.
135
11
136
137
Figure S7. A selected EvoMining Prediction. This BGC was predicted after identification of a
138
recruited AroA homolog which was not identified by ClusterFinder or antiSMASH. Detailed
139
annotation is available as table S7.
12