CINF 4: Naming algorithms for derivatives of peptide-like natural products

Naming algorithms for
derivatives of peptide-like
natural products
Roger Sayle& Noel O’Boyle
Nextmove software, cambridge, uk
Christopher southan,
Iuphar/bps guide to pharmacology
250th ACS National Meeting, Boston, MA. Sunday 16th August 2015

30 second overview
• This talk describes the development of software for
both naming peptides and reading peptide names
matching the de facto standard practices currently
followed by biochemists.
• Unlike computer representations, like Pisotoia HELM
or InChI keys, these names and identifiers match
those typically found in the scientific literature and
vendor catalogues.
• A significant application of this technology is to check
and correct peptide representations in databases.

Problem motivation
• Maintaining a database of biological activities of
mature proteins and peptides presents a significant
technical challenge.
• IUPHAR/BPS’ “Guide to Pharmacology” and EBI’s
ChEMBL represent current state-of-the-art efforts to
capture/represent peptide-like ligands.
• The ligands require more than (FASTA) bioinformatics
including disulfide bridging architecture, non-
standard amino acids, post-translational
modifications, N- and C- terminal modifications etc.

Iuphar/bps gtop peptides
• The “Guide to Pharmacology” database contains:
– The common name “oxytocin”
– The species, e.g. “human”
– The UniProt ID “P001178”
– The 1-letter Sequence “CYIQNCPLG”
– The 3-letter sequence “Cys-Tyr-Ile-…-Leu-Gly-NH2”
– A text description “Post-translation modification”, e.g. “A
disulfide bond is formed between cysteine residues at
positions 1 and 5 and the C-terminal glycine is amidated”.
– Often SMILES and standard InChIKey.

Problem 1: consistency
• The challenge with these advanced formats are that
the names, three-letter codes and modification
descriptions are text-locked, unreadable by software.
Examples of errors and inconsistency:
Ligand #4463: “PheGlnThrSerGluAlaIleLeuPro…”
Ligand #1335: “…Leu-Arg-AlaPro-Leu-Lys...”
Ligand #8263: “val-leu-gln-glu-leu-asn-val-thr-val”
Ligand #5873: “…Pr-oGl-yGl-ySe-rMe-tLy-sLe-u…”
Ligand #3591: “PHQLLRVPro-His-Ala-Gln-Leu…”

Problem 2: ambiguity
• Ligand #3630 (neuropeptide B29) “(Br)Trp” with note
“The n-terminal tryptophan is brominated”.
– Suggested replacement Trp(6-Br)
• In Ligand #1036, “(Ac)Ala” means N2-acetyl but
“(Ac)Lys” means N6-acetyl, in #1188 “Ac-” appears
without parenthesis, in Ligand #3853, “AcPhe-”
appears without a hyphen…
– Suggest Ac- at N-term, -N(Ac)Phe infix, Lys(Ac) sidechain
– This even allows Ac-N(Ac)Lys(Ac)-OH, aka. N(Ac2)Lys(Ac).

Problem 3: disulfide bridging
• Capturing the disulfide bridging architecture in the
three-letter (condensed) representation allows it to
be read/checked for errors.
• This is done in some places but not in others.
• Disulfide bridges are particularly tricky even for the
folks at UNIPROT: Annexin I (ligand #1031, P04083)
isn’t annotated as disulfide bridged, despite the 3D
structure in PDB 1HM6, and the experimental
evidence of an intramolecular disulfide described in
PubMed 7663390.

Types of peptide name/identifier
• Sequence: CYIQDCPLG
• Peptide Name: [Asp5]oxytocin or [5-L-aspartic acid]oxytocin
• Chemical IUPAC Name:
2-[(4R,7S,10S,13S,16S,19R)-19-amino-4-[(2S)-2-[[(1S)-1-[(2-amino-2-oxo-ethyl)carbamoyl]-3-methyl-
butyl]carbamoyl]pyrrolidine-1-carbonyl]-10-(3-amino-3-oxo-propyl)-16-[(4-hydroxyphenyl)methyl]-13-
[(1S)-1-methylpropyl]-6,9,12,15,18-pentaoxo-1,2-dithia-5,8,11,14,17-pentazacycloicos-7-yl]acetic acid
• Biological IUPAC Name
L-cysteinyl-L-tyrosyl-L-isoleucyl-L-glutaminyl-L-alpha-aspartyl-
L-cysteinyl-L-prolyl-L-leucyl-glycinamide (1->6)-disulfide
• Condensed: Cys(1)-Tyr-Ile-Gln-Asp-Cys(1)-Pro-Leu-Gly-NH2
• Pistoia HELM:
PEPTIDE1{C.Y.I.Q.N.C.P.L.G.[am]}$PEPTIDE1,PEPTIDE1,1:R3-6:R3$$$

Helm teething problems
• Pistoia’s HELM notation marks a significant advance
over the limitations of one-letter bioinformatics.
• Alas, its original goals didn’t include data exchange,
which has only recently been addressed by the
extensions of inlineHELM and XHELM [and fixes from
NextMove Software for improved interoperability].
• Alas, this still doesn’t address some core limitations:
– Pistoia Monomer Library: PEPTIDE1{[fmoc].A}$$$$
– EBI ChEMBL Monomers: PEPTIDE1{[Fmoc_A]}$$$$

Iupac condensed names (chembl)
• The following names are machine generated
• H-Cys-Pro-Trp-His-Leu-Leu-Pro-Phe-Cys-OH CHEMBL501567
• H-Tyr-Pro-Phe-Phe-OtBu CHEMBL500195
• cyclo[Ala-Tyr-Val-Orn-Leu-D-Phe-Pro-Phe-D-Phe-Asn] CHEMBL438006
• H-Nle(Et)-Tyr-Pro-Trp-Phe-NH2 CHEMBL500704
• H-DL-hPhe-Val-Met-Tyr(PO3H2)-Asn-Leu-Gly-Glu-OH CHEMBL439086
• cyclo[Phe-D-Trp-Tyr(Me)-D-Pro] CHEMBL507127
• H-D-Pyr-D-Leu-pyrrolidide CHEMBL1181307
• Ac-DL-Phe-aThr-Leu-Asp-Ala-Asp-DL-Phe(4-Cl)-OH CHEMBL1791047
• H-D-Cys(1)-D-Asp-Gly-Tyr(3-NO2)-Gly-Hyp-Asp-D-Cys(1)-NH2 CHEMBL583516
• Boc-Tyr-Tyr(3-Br)-OMe CHEMBL1976073

peptide names (chembl)
• The following names are machine generated
• [15-L-arginine]nociceptin CHEMBL526333
• [2-4-chloro-L-phenylalanine]neuropeptide S [human] CHEMBL441576
• [1-L-threonine]cyclosporin A CHEMBL2370014
• [6-L-tryptophan]sermorelin free acid CHEMBL440438
• angiotensin II (3-8) CHEMBL261120
• nociceptin amide CHEMBL389521
• acetyl-alpha-MSH (4-10) amide CHEMBL410411
• [2-L-cysteine,13-L-cysteine]neurotensin disulfide CHEMBL3278512
• myristoyl-[1-L-lysine,4-L-tryptophan]tetrapandin 2 amide CHEMBL3288219
• [2-(4RS)-thiazolidine-4-carboxylic acid,4-L-proline]endomorphin-2 CHEMBL126611
• [22-L-serine]kalata B1 CHEMBL1801140

Advanced Peptide names
• Named peptides imply not only sequence but also
N-terminal acetylation, C-terminal amidation and
disulfide bridge topology.
• Example derivative naming operations:
– gastrin (14-17)
– motilin amide
– oxytocin free-acid
– acetyl-oxytocin
– deacetyl-abarelix
– oxytocin reduced
– endothelin-1 (1→3),(11 → 15)-bis(disulfide)

homodetic cycles #1
cyclo[Leu-D-Phe-Pro-Val-Orn-Leu-D-Phe-Pro-Val-Orn]
gramicidin S

homodetic cycles #2
cyclo[OAla-Val-D-OVal-D-Val-OAla-Val-D-OVal-D-Val-OAla-Val-D-OVal-D-Val]
valinomycin

Ambiguous/Preferred forms
• [3-L-isoleucine]lypressin vs. [8-L-lysine]vasotocin
• [2-L-phenylalanine]lypressin vs. [8-L-lysine]phenypresin
• [2-L-phenylalanine]ornipressin vs. [8-L-ornithine]phenypressin
• [3-L-isoleucine]ornipressin vs. [8-L-ornithine]vasotocin
• [4-L-methionine]afamelanotide vs. [7-D-phenylalanine]α-MSH
• [Thr1,Lys2]endomorphin-1 vs. [Trp3,Phe4]tuftsin amide
• [Gln3]thyrotropin-releasing hormone vs. [Pro3]eisenin amide
• [Trp1,Val2]endomorphin-2 vs. [Val2,Phe3]gastrin tetrapeptide

Named cyclic peptide derivatives
• Mutants of named cyclic peptides are identified by
comparing against all “rotational” permutations.
Example line notation query (CHEMBL478596)
cyclo[Ala-Gly-Thr-Phe-Val-Tyr]
Reference database line notations:
cyclo[Gly-Thr-Phe-Leu-Tyr-Thr] dichotomin B
cyclo[Ala-Gly-Thr-Phe-Leu-Tyr] dichotomin C
Resulting Sugar & Splice peptide name:
[5-L-valine]dichotomin C

Lower locants in cyclic peptides
• Symmetric cyclic peptides provide an interesting
challenge, where substitions are different locants can
potentially be synonymous.
• CHEMBL1934531
– [3-(4S)-4-amino-L-proline]gramicidin S preferred
– [8-(4S)-4-amino-L-proline]gramicidin S acceptable
• CHEMBL1934536
– [3-(4R)-4-amino-L-proline]gramicidin S preferred
– [8-(4R)-4-amino-L-proline]gramicidin S acceptable

Scaling-up protein variant naming
• The algorithm described for naming peptides can
also be applied to naming arbitary protein variants.
• Consider the a database of the following 11 peptides:
– CFFQNCPRG phenylpressin
– CFVRNCPTG annetocin
– CFWTSCPIG octopressin
– CYFQNCPRG argipressin
– CYFQNCPKG lypressin
– CYFRNCPIG cephalotocin
– CYIQNCPLG oxytocin
– CYIQNCPPG prol-oxytocin
– CYIQNCPRG vasotocin
– CYIQSCPIG seritocin
– CYISNCPIG isotocin

Dag representation of sequences
These 11 peptides may be efficiently represented and
search as a “directed acyclic graph” [38 vs. 99 states]

entirety of uniprot/swissprot
• Using this representation, all 540546 protein
sequences in uniprot_sprot, which contains over
192M amino acids, requires 142M states (1.4Gb).
• This data structure allows close analogues to be
identified much faster than using NCBI blastp.
• For example, all 540546 sequences can be queried
against this database (i.e. all-against-all) in ~9m30s
on a single core on a laptop.
• The sequence from PDB 1CRN (crambin 46AA) is
canonically named as [L25I]P01542 in 0.002s.

Application to precision medicine
• A more realistic example is that sequence of the
gene “spastic paraplegia4” with six mutations from
OMIM:604277 can be canonically named as
[I344K,S362C,N386S,D441G,C448Y,R499C]Q9UBP0
• Run-time for this query is 0.2s.
• By comparison, blastp 2.2.29+ takes about 6s.
– With default arguments, NCBI blastp run time is 7s.
– Only 6s with –num_descriptions 1 –num_alignments 1.

30 second summary
• This talk describes the development of software for
both naming peptides and reading peptide names
matching the de facto standard practices currently
followed by biochemists.
• Unlike computer representations, like Pisotoia HELM
or InChI keys, these names and identifiers match
those typically found in the scientific literature and
vendor catalogues.
• A significant application of this technology is to check
and correct peptide representations in databases.

acknowledgements
• Joanna Sharman, IUPhar, Edinburgh University, UK.
• Lisa Sach-Peltason, Hoffmann-La Roche, Basel.
• Joann Prescott-Roy, Novartis, Boston, MA.
• Greg Landrum, Novatis, Basel, Switzerland.
• Evan Bolton, NCBI PubChem project, Bethesda, MD.
• Daniel Lowe, NextMove Software, Cambridge, UK.
• John May, NextMove Software, Cambridge, UK.

CINF 4: Naming algorithms for derivatives of peptide-like natural products

More Related Content

What's hot (12)

Viewers also liked (20)

Similar to CINF 4: Naming algorithms for derivatives of peptide-like natural products (20)

More from NextMove Software (19)

Recently uploaded (20)

CINF 4: Naming algorithms for derivatives of peptide-like natural products