Bioinformatics analysis of S. cerevisiae NIT1 (P40447 / YIL164C)

Reproducible pipeline: just all (see justfile). All scripts read their inputs
from data/ at run time; no results are hardcoded in code.

Question

NIT1 (P40447, YIL164C) is a 199-aa putative nitrilase-superfamily protein (carbon-
nitrogen hydrolase superfamily; Pfam PF00795 CN_hydrolase; PANTHER PTHR46044:SF1).
Three questions were addressed:

1. Is the nitrilase-superfamily Glu-Lys-Cys (E-K-C) catalytic triad intact?
2. Is NIT1 a truncated / split ORF (UniProt CAUTION: "Could be the product of a
   pseudogene. NIT1/YIL164C seems to be the N-terminal part of a putative nitrilase-
   like protein formed of NIT1/YIL164C and YIL165C")?
3. Orthology to human NIT1 / NIT2.

Data

Sequences fetched from UniProt REST (data/*.fasta).

Methods

• scripts/catalytic_triad.py — reports the residue at each given 1-based position
  in a FASTA (positions supplied on the command line; not hardcoded).
• MAFFT --auto alignment of the five sequences (results/combined.aln.fasta).
• scripts/analyze_alignment.py — maps the reference's PROSITE-predicted triad
  positions to alignment columns and reports what every other sequence has in
  those columns, plus each sequence's non-gap span (truncation proxy).
• scripts/test_split.py — builds the artificial NIT1+YIL165C concatenation and
  aligns it with the real family members to test the split-ORF hypothesis.
• scripts/pairwise_identity.py — % identity of NIT1 to each homolog over
  co-aligned (gap-free) columns.

Results

1. Catalytic triad is present in NIT1

PROSITE-predicted active sites map to E44, K135, C169 in NIT1
(results/nit1_triad.tsv):

All three canonical nitrilase-superfamily triad residues are present at the
predicted positions. In the multiple alignment these three residues fall in the
same columns (87, 188, 232) occupied by the E-K-C triad of human NIT1, human NIT2
and Arabidopsis NIT1 (results/alignment_analysis.txt) — i.e. the triad is
conserved and correctly positioned. The nucleophilic Cys sits in the
family-diagnostic GGAICWEN motif.

Positive control (results/nit2_triad_control.tsv): running the same script on
human NIT2 at its own UniProt active-site positions (E43, K112, C153) correctly
returns E/K/C, confirming the script reads residues from whatever
sequence/positions it is given rather than returning a fixed answer.

2. NIT1 is the N-terminal ~2/3 of a nitrilase fold; YIL165C is the C-terminal ~1/3

Non-gap alignment spans (results/split_test.txt, results/alignment_analysis.txt):

In the split-test alignment, NIT1 (cols 1–270) and YIL165C (cols 271–406) tile
the domain almost exactly end-to-end with no overlap, and the 318-aa
concatenation spans the full 1–406 columns — the same span as the full-length
family members (human NIT1 327 aa; Arabidopsis 346 aa). The complete E-K-C triad
lies entirely within the NIT1 fragment (cols 87/188/232, all < 270); YIL165C
contributes only the C-terminal region downstream of the triad and carries none of
the triad residues (-/-/- at the triad columns).

Interpretation. This is strong sequence-level support for the UniProt CAUTION:
YIL164C + YIL165C together reconstitute one complete nitrilase-superfamily domain
of normal length, split across two adjacent annotated ORFs. NIT1 alone, although it
retains all three catalytic residues, is missing the C-terminal ~1/3 of the
fold. In characterised nitrilase-superfamily enzymes the C-terminal region
contributes to the α-β-β-α sandwich fold and to the oligomerisation interfaces
required for activity, so a fragment ending at the equivalent of column ~270 is very
unlikely to fold into an active enzyme on its own. Whether YIL164C+YIL165C are
translated as one protein (e.g. via a sequencing error, frameshift, or read-through
that the reference genome splits) or are a genuinely disrupted pseudogene cannot be
resolved from sequence alone.

3. Orthology / identity

Percent identity of NIT1 over co-aligned columns (results/pairwise_identity.tsv):

NIT1 shows the expected carbon-nitrogen-hydrolase-superfamily level of similarity to
all three reference enzymes (~25–47% over the aligned catalytic core). It is not
markedly closer to human NIT1 than to human NIT2, and both yeast ORFs sit in the
uncharacterised PANTHER subfamily PTHR46044:SF1 "CN hydrolase domain-containing
protein" — i.e. not in the substrate-defined subfamilies of the same family
(SF14 arylacetonitrilase, SF4 cyanide hydratase, SF11 nitrilase-1-related). A clean
1:1 orthology assignment to a specific human paralog is therefore not supported;
NIT1 is a divergent fungal member of the superfamily whose substrate-specific
subfamily is undefined.

Bottom line

• The E-K-C catalytic triad residues are individually intact in NIT1 (E44, K135,
  C169), conserved and correctly positioned relative to characterised family members
  — so NIT1 is not a triad-loss pseudoenzyme.
• However, NIT1 encodes only the N-terminal ~2/3 of a nitrilase-superfamily
  domain; the C-terminal third is separately annotated as YIL165C, and the two
  concatenate to a full-length domain. NIT1 as annotated is therefore a truncated
  fragment unlikely to be an active enzyme on its own, consistent with the UniProt
  pseudogene CAUTION and its PE=5 ("Uncertain") status.
• No specific catalytic substrate can be assigned. The intact triad argues the
  ancestral fold was catalytic, but membership in the uncharacterised PTHR46044:SF1
  subfamily (not a substrate-defined branch) and the truncation mean the physiological
  activity — if any — is unknown. A generic "nitrilase activity" assignment is not
  supported by these data.

Checklist

• [x] Scripts use no hardcoded inputs/outputs (all positions/paths passed as args;
  concatenation computed at run time).
• [x] Triad script tested on a second, unrelated input (human NIT2 positive control
  returns the correct E/K/C).
• [x] Analyses completed as expected (MAFFT + all scripts ran; outputs in results/).
• [x] Direct script outputs are in results/.
• [x] Summary includes provenance and justification; uncertainty explicitly stated
  where sequence alone cannot resolve the question (translation vs pseudogene).