Definition of chorismate biosynthesis

GapMind for Amino acid biosynthesis

Definition of chorismate biosynthesis

# Chorismate is the starting point for the biosynthesis of the aromatic amino acids
# phenylalanine, tryptophan, and tyrosine.
# Chorismate biosynthesis in GapMind is based on 
# MetaCyc pathways chorismate biosynthesis I (metacyc:ARO-PWY), from
# D-erythrose-4-phosphate and phosphoenolpyruvate, or
# II (metacyc:PWY-6165), from D-glyceraldeyde-3-phosphate and L-asparatate.
# Both pathways are identical after they reach 3-dehydroquinate.

# This is also known as DAHP (3-deoxy-D-arabino-heptulosonate 7-phosphate) synthase.
# Add CA265_RS11635 (a fusion with chorismate mutase) because it is diverged,
# is confirmed by cofitness, and is essential in other Bacteroidetes.
# uniprot:P39912 (Bacillus subtilis aroA) is annotated in BRENDA as chorismate mutase but it also has this activity
# (PMC1198938).
# CH_123440 is annotated as this but without the EC number.
# 4.1.2.15 is an obsolete EC number, but it appears in a few entries, so is ignored.
# uniprot:C9K7C8 is annotated as this in SwissProt but its function is uncertain, so it is ignored.
aroG	3-deoxy-7-phosphoheptulonate synthase	EC:2.5.1.54	curated:BRENDA::P39912	curated:CharProtDB::CH_123440	curated:reanno::Pedo557:CA265_RS11635	ignore_other:EC 4.1.2.15	ignore:SwissProt::C9K7C8

# (There's also the obsolete EC number 4.6.1.3, no longer used.)
aroB	3-dehydroquinate synthase	EC:4.2.3.4

# uniprot:B9CK59 may be misannotated in BRENDA. Several plant shikimate dehydrogenases may also be
# 3-dehydroquinate dehydratases (fusion proteins), so any similarity to shikimate dehydrogenase
# (EC 1.1.1.24 or 1.1.1.25 or 1.1.1.282) is ignored.
# PYRFU_RS04235 (uniprot:G0EDV3) from Pyrolobus fumarii is diverged, especially at the C terminal part,
# but has conserved functional residues, and a homolog from Sedimentisphaera salicampi is fused to
# aroE, so it is probably the missing aroD.
aroD	3-dehydroquinate dehydratase	EC:4.2.1.10	ignore:BRENDA::B9CK59	ignore_other:EC 1.1.1.24	ignore_other:EC 1.1.1.25	ignore_other:EC 1.1.1.282	predicted:G0EDV3

# EC:1.1.1.282 is with NAD(P)H instead of NADPH.
# BT4215 from Bacteroides thetaiotaomicron (uniprot:Q8A006_BACTN) is diverged, is the only good candidate,
# and is essential in various Bacteroidetes.
# Ga0059261_2194 / BDW16_RS10815 (uniprot:A0A2M8WD96) has auxotrophic phenotypes in RB-TnSeq data and
# can complement an aroE- strain of E. coli (Bradley Biggs).
# CH_122204 in CharProtDB is ignored because it is probably quinate dehydrogenase, not shikimate dehydrogenase.
aroE	shikimate dehydrogenase	EC:1.1.1.25	EC:1.1.1.282	uniprot:Q8A006_BACTN	uniprot:A0A2M8WD96	ignore:CharProtDB::CH_122204

# In E. coli, AroL and AroK are isozymes.
# In Bacillus subtilis, this gene was known as AroI, and it was cloned by complementation
# (see A. Nakane et al, J. Fermentation and Bioengineering 1994, 77:312-314.)
# That sequence is identical to uniprot:AROK_BACSU.
# Manually add BT3393 (uniprot:AROK_BACTN) from B. thetaiotaomicron because it is diverged, is the only good candidate,
# and is essential in various Bacteroidetes.
# Similarly for Echvi_0140 (uniprot:L0FT15_ECHVK) from Echinicola vietnamensis.
# And DVU0892 (uniprot:AROK_DESVH) from D. vulgaris Hildenborough is confirmed by cofitness
# CCNA_03103 (uniprot:AROK_CAUVN) is confirmed by cofitness and similar proteins
# such as SMc00695 (uniprot:AROK_RHIME) and PGA1_c14090 (uniprot:A0A135IJ25) are essential.
aroL	shikimate kinase	EC:2.7.1.71	uniprot:AROK_BACSU	uniprot:AROK_BACTN	uniprot:L0FT15_ECHVK	uniprot:AROK_DESVH	uniprot:AROK_CAUVN	uniprot:AROK_RHIME	uniprot:A0A135IJ25

# Add AroA from Desulfovibrio vulgaris (DVU0463) because it is a bit diverged, is conserved essential,
# and clusters with aromatic amino acid biosynthesis genes.
# HMPREF1058_RS13970 (uniprot:I9A2E3) is cofit with chorismate synthase (Surya Tripathi); also, it is 71% identical to
# BT2186 / BT_RS11065, which can complement an aroA- strain of E. coli (Bradley Biggs).
# PYRFU_RS00635 (uniprot:G0EEF0) from Pyrolobus fumarii is from this family, has similar active site residues
# (alignment to Q83E11 shows three conserved residues, and D315 => R331), and
# its closest homologs are in chorisomate synthesis operons (but, these are under 30% identity);
# it is probably the missing aroA.
aroA	3-phosphoshikimate 1-carboxyvinyltransferase	EC:2.5.1.19	uniprot:Q72EV5_DESVH	uniprot:I9A2E3	uniprot:L0FR45	predicted:G0EEF0

# (There's also obsolete EC 4.6.1.4, no longer used)
aroC	chorismate synthase	EC:4.2.3.5

# The triose-phosphate isomerase tpiA is also thought to convert D-glyceraldehyde 3-phosphate to enolaldehyde, which
# spontaneously converts to methylglyoxal.
# (Alternatively, methylglyoxal might be formed by methylgyoxal synthase, EC 4.2.3.3?)
# Ignore uniprot:P00941, which is misannotated in BRENDA.
tpiA	D-glyceraldehyde-3-phosphate phospholyase	EC:5.3.1.1	ignore:BRENDA::P00941

# 6-deoxy-5-ketofructose-1-phosphate synthase is an activity of some fructose-bisphosphate aldolases
# (which are usually annotated as 4.1.2.13). To find the fbp in Desulfovibrio vulgaris Hildenborough
# and Miyazaki F, it is necessary to match more broadly.
# MetaCyc reports that AroA' from Methanococcus jannaschii (metacyc:MONOMER-14592)
# also has activity as a fructose-bisphosphate aldolase, but it's not clear that
# it carries out this reaction.
# The bifunctional fructose-1,6-bisphosphate aldolase/phosphatases are ignored because it
# is not obvious that they would carry out this reaction; this includes
# uniprot:Q980K6, uniprot:A3MSD2, uniprot:Q8NKR9, uniprot:A0RV30, uniprot:A4YIZ5, uniprot:A8A9E4,
# uniprot:B1YAL1, uniprot:B6YTP6, uniprot:D9PUH5, uniprot:F9VMT6, uniprot:Q2RG86, uniprot:Q72K02.
# And ignore CharProtDB items with incorrect EC and very short SwissProt entries.
fbp	6-deoxy-5-ketofructose 1-phosphate synthase	EC:2.2.1.11	EC:4.1.2.13	ignore:metacyc::MONOMER-14592	ignore_other:fructose%bisphosphate aldolase	ignore:SwissProt::P84722	ignore:SwissProt::P86979	ignore:SwissProt::P86980	ignore:BRENDA::Q980K6	ignore:BRENDA::A3MSD2	ignore:SwissProt::Q8NKR9	ignore:SwissProt::A0RV30	ignore:SwissProt::A4YIZ5	ignore:SwissProt::A8A9E4	ignore:SwissProt::B1YAL1	ignore:SwissProt::B6YTP6	ignore:SwissProt::D9PUH5	ignore:SwissProt::F9VMT6	ignore:SwissProt::Q2RG86	ignore:SwissProt::Q72K02

import met.steps:aspartate-semialdehyde

# aroA' condenses 6-deoxy-5-ketofructose 1-phosphate with L-aspartate 4-semialdehyde
aroA'	2-amino-3,7-dideoxy-D-threo-hept-6-ulosonate synthase	EC:2.2.1.10

# Ignore uniprot:P81230, which is misannotated as this in BRENDA
aroB'	dehydroquinate synthase II	EC:1.4.1.24	ignore:BRENDA::P81230

# Pathway I uses aroG and aroB, while pathway II uses non-canonical activities of triose-phosphate
# isomerase (tpiA) and fructose-bisphosphate aldolase (fbp)
# to form 6-deoxy-5-ketofructose 1-phosphate. AroA' condenses this with asparate semialdehyde
# to 2-amino-3,7-dideoxy-D-threo-hept-6-ulosonate,
# and AroB' cyclizes it to 3-dehydroquinate.
3-dehydroquinate: aroG aroB
3-dehydroquinate: tpiA fbp aspartate-semialdehyde aroA' aroB'

all: 3-dehydroquinate aroD aroE aroL aroA aroC

Downloads

Candidates (tab-delimited)
Steps (tab-delimited)
Rules (tab-delimited)
Protein sequences (fasta format)
Organisms (tab-delimited)
SQLite3 databases

Related tools

About GapMind

Each pathway is defined by a set of rules based on individual steps or genes. Candidates for each step are identified by using ublast (a fast alternative to protein BLAST) against a database of manually-curated proteins (most of which are experimentally characterized) or by using HMMer with enzyme models (usually from TIGRFam). Ublast hits may be split across two different proteins.

A candidate for a step is "high confidence" if either:

ublast finds a hit to a characterized protein at above 40% identity and 80% coverage, and bits >= other bits+10.
- (Hits to curated proteins without experimental data as to their function are never considered high confidence.)
HMMer finds a hit with 80% coverage of the model, and either other identity < 40 or other coverage < 0.75.

where "other" refers to the best ublast hit to a sequence that is not annotated as performing this step (and is not "ignored").

Otherwise, a candidate is "medium confidence" if either:

ublast finds a hit at above 40% identity and 70% coverage (ignoring otherBits).
ublast finds a hit at above 30% identity and 80% coverage, and bits >= other bits.
HMMer finds a hit (regardless of coverage or other bits).

Other blast hits with at least 50% coverage are "low confidence."

Steps with no high- or medium-confidence candidates may be considered "gaps." For the typical bacterium that can make all 20 amino acids, there are 1-2 gaps in amino acid biosynthesis pathways. For diverse bacteria and archaea that can utilize a carbon source, there is a complete high-confidence catabolic pathway (including a transporter) just 38% of the time, and there is a complete medium-confidence pathway 63% of the time. Gaps may be due to:

our ignorance of proteins' functions,
omissions in the gene models,
frame-shift errors in the genome sequence, or
the organism lacks the pathway.

GapMind relies on the predicted proteins in the genome and does not search the six-frame translation. In most cases, you can search the six-frame translation by clicking on links to Curated BLAST for each step definition (in the per-step page).

For more information, see:

the paper from 2019 on GapMind for amino acid biosynthesis
the paper from 2022 on GapMind for carbon sources
the source code
instructions for running GapMind on your computer
changes to Amino acid biosynthesis since the publication.

If you notice any errors or omissions in the step descriptions, or any questionable results, please let us know

by Morgan Price, Arkin group, Lawrence Berkeley National Laboratory

GapMind for Amino acid biosynthesis