mitfa / mitfb

Back to pairs

Bottom line: PARTITION (subfunctionalization) at the expression level, with some
overlap still left. The ancestral Mitf expression domains are split: mitfa alone is
expressed in neural crest melanoblasts, mitfb alone in the epiphysis and olfactory bulb, and
both in the RPE. The two proteins are still interchangeable in melanophores, since mitfb
rescues nacre/mitfa when driven by the mitfa promoter. The overlap shows up as partial
redundancy in xanthophores and, together with tfec, in RPE pigmentation. Medaka resolved the
same pair differently, which points to lineage-specific resolution of the duplicates.

	mitfa	mitfb
UniProt	Q9PWC2	F1Q885
Human ortholog	MITF	MITF
Review	genes/DANRE/mitfa	genes/DANRE/mitfb

1. Evidence the pair comes from the TGD

PANTHER (row from panther_tgd_pairs.tsv, family PTHR45776):

tgd_call	duplication_branch	pair_class	gar_coorthologs	medaka_coorthologs	medaka_parallel_dup	human (a / b)
TGD_likely_parallel	Teleostei|DANRE	1:1	1	2	yes	MITF(LDO) / MITF(O)

PANTHER places the duplication on the zebrafish terminal branch, not on the TGD branch. It
does not group the zebrafish and medaka copies as ((zfA, medA), (zfB, medB)). Both zebrafish
copies share one spotted gar co-ortholog, and medaka also has two copies. The project page
explains why this pattern is expected for a TGD pair with lineage-specific resolution, or a
poorly resolved tree. The tree alone therefore does not establish a TGD origin.

Synteny and comparative genomics give independent support:

• Two mitf genes are present across teleosts:
  [PMID:12019239 "Two separate genes (mitf-m and mitf-b), however, are present in different teleost fish species including the poeciliid Xiphophorus, the pufferfishes Fugu rubripes and Tetraodon nigroviridis, and the zebrafish Danio rerio."]
  They are dated to
  [PMID:12019239 "duplication at least 100 million years ago"].
• Braasch et al. list the pair among fish-specific genome duplication (FSGD) duplicates. The
  paralogs sit on different linkage groups, and each has an ortholog in medaka, stickleback
  and Tetraodon (zebrafish chr 6 and chr 23; medaka 5 and 7). Their table text reads:
  [PMID:20333216 "MitfaHsa3mitfaDre6Ola5GacXVIITni11lRef. 13mitfbDre23Ola7GacXIITni9"].
• Medaka has both copies:
  [PMID:37823232 "We found two mitf paralogous genes, mitfa and mitfb, in the medaka genome, as in zebrafish"].

Established: a pre-euteleost origin shared by ostariophysans and acanthopterygians, with
conserved synteny across four teleost genomes. This is consistent with the TGD, and the pair
is treated as TGD ohnologs in the literature. Not established here: a gene-tree node at
the base of teleosts. PANTHER does not recover one, and no gar-bridge synteny analysis was
found for this pair.

2. Protein-level comparison

• Identity: 55.0% identity and 64.7% similarity over 515 alignment columns
  (annotation-comparison.md). mitfa is 412 aa and mitfb is
  500 aa. Both have the same InterPro/Pfam architecture: MiT/TFE N-terminal region, bHLH-LZ,
  and MiT/TFE C-terminal region. mitfb has an N-terminal extension (UniProt).
• Isoform correspondence: the N-termini differ like the alternative first exons of
  tetrapod MITF:
  [PMID:11543618 "The proteins encoded by the two zebrafish mitf genes appear homologous to distinct isoforms generated by alternately spliced mRNAs of the single mammalian Mitf gene"];
  mitfb has
  [PMID:23139843 "greater similarity, especially in the amino terminal region, to Mitf isoforms present in the mouse RPE"].
  Across fish,
  [PMID:12019239 "Fish proteins MITF-m and MITF-b correspond at both the structural and the expression levels to one particular bird/mammalian MITF isoform."]
  Exon degeneration left traces:
  [PMID:12019239 "a remnant of the first exon specific for MITF-m is detected within the pufferfish gene encoding MITF-b"].
• Cross-rescue (key result):
  [PMID:11543618 "mitfb, but not a zebrafish ortholog of the closely related gene tfe3, can rescue neural crest melanophore development in nacre/mitfa mutant embryos when expressed via the mitfa promoter"].
  Both copies therefore keep the ancestral melanogenic DNA-binding transcription factor
  activity. The reciprocal test (mitfa protein driven by an mitfb promoter) has not been
  reported.
• Medaka contrast (inference only): Li et al. describe medaka duplicates as mitf1/mitf2
  [PMID:24118994 "The fish medaka has two mitf genes, mitf1 and mitf2."]
  and report protein-level degeneration of one copy:
  [PMID:23781143 "Compared to mitf1, mitf2 exhibits an accelerated sequence divergence and loses melanocytic expression in embryos at critical developmental stages."]
  How mitf1/mitf2 map onto mitfa/mitfb is not stated in the cached abstracts, so this is not
  used to infer divergence of zebrafish mitfb.

Conclusion: the core molecular function is conserved in both copies. The protein-level
difference is the N-terminal isoform-type region, which may affect target specificity; this
has not been tested.

3. Expression

• Tetrapod single-copy state:
  [PMID:23139843 "The mammalian Mitf gene produces numerous isoforms through the use of at least nine different alternative promoters and alternative splicing"].
  Melanocyte (M) and RPE (A/H) isoforms come from different promoters of one gene.
• mitfa: expressed in all embryonic melanogenic cells:
  [PMID:11543618 "Mutations in the zebrafish nacre/mitfa gene, expressed in all embryonic melanogenic cells, perturb only neural crest melanocytes"];
  [PMID:23139843 "However, mitfa is expressed in the RPE as well as neural crest cells of the developing zebrafish embryo."]
• mitfb: RPE, but not neural crest:
  [PMID:11543618 "mitfb is coexpressed with mitfa in the RPE at an appropriate time to compensate for loss of mitfa function in the nacre mutant but is not expressed in neural crest melanoblasts"];
  plus copy-specific domains:
  [PMID:11543618 "Additionally, mitfb is expressed in the epiphysis and olfactory bulb where mitfa is not, and where Mitf expression has not previously been reported in other species."]
  mitfb expression in xanthophores is equivocal:
  [PMID:40123122 "mitfb transcripts are detectable in xanthophores by some but not other analyses"].
• Ancestral (gar) state: no published gar mitf expression pattern was found. Whether the
  epiphysis and olfactory bulb domains are ancestral (lost from tetrapods, or not yet seen
  there) or new in teleosts is unknown.

Domain	mitfa	mitfb	Tetrapod MITF
Neural crest melanoblasts / melanophores	yes	no	yes (M isoform)
RPE	yes	yes	yes (A/H isoforms)
Epiphysis, olfactory bulb	no	yes	not reported (per PMID:11543618)

4. Experimental evidence of function

mitfa
- nacre (w2) null: no neural crest melanophores, normal RPE:
[PMID:10433906 "Homozygous nacre (nac(w2)) mutants lack melanophores throughout development but have increased numbers of iridophores."];
[PMID:10433906 "The non-crest-derived retinal pigment epithelium is normal"].
- Allele type: w2 is a premature stop codon (PTC)
[PMID:23139843 "First, three of the four alleles in question (mitfaw2, mitfbhu3857 and mitfbhu3561) create premature stop codons"].
An RNA-less deletion allele rules out transcriptional adaptation for mitfa:
[PMID:40123122 "showing that mitfa alone is not essential for xanthophore development and making transcriptional adaptation an unlikely explanation for xanthophore persistence in the premature termination alleles"].
- Required at several stages of differentiation:
[PMID:21146516 "mitfa is required at several stages of melanocyte differentiation"].
- Sufficient to pigment the retina when misexpressed:
[PMID:23139843 "the ability of mitfa to induce pigmentation in the zebrafish retina when misexpressed"].
- Redundant with tfec for choroid fissure closure:
[PMID:32541011 "mitfa;tfec mutants possess severe colobomas"].

mitfb
- ENU nonsense alleles hu3857 (exon 2) and hu3561 (exon 7, bHLH-LZ), both PTC:
[PMID:23139843 "no obvious abnormalities in homozygous mitfb embryos or larvae were observed"].
An RNA-less mitfb allele has not been reported, so masking by transcriptional
adaptation [PMID:30944477 "mutant alleles with minimal transcriptional adaptation-derived compensation."]
is not excluded for mitfb.

Double mutants and compensation
- RPE: [PMID:23139843 "the loss of Mitf activity in mitfa, mitfb, or double mitf mutant zebrafish had no effect on RPE pigmentation or development"].
Adding tfec loss does affect it:
[PMID:42663426 "simultaneous null mutations in all three genes cause a complete loss of all four chromatophore types and lead to defective pigmentation of the retinal pigmented epithelium (RPE)"]
(abstract; the RPE claim is stated for Oryzias and zebrafish).
- Xanthophores, two independent labs:
[PMID:37823232 "Compared with a single mitfa mutant, mitfa−/−; mitfb−/− double mutants showed a slightly more severe reduction of xanthophores, with fewer fluorescent xanthophores on the trunk"];
[PMID:40123122 "These phenotypes suggest that mitfb partially compensates for the loss of mitfa (though not tfec) in xanthophores and especially so in the fins."]
- Cross-rescue: see section 2 (mitfb rescues nacre melanophores).

Medaka (lineage comparison)
- [PMID:37823232 "Single loss of mitfb resulted in a defect in melanophore formation, being delayed at 2 days post-fertilization (dpf)"];
[PMID:37823232 "The double loss of mitfa and mitfb resulted in complete absence of not only melanophores but also xanthophores and leucophores throughout embryogenesis"];
[PMID:37823232 "suggesting that mitfb, but not mitfa, plays the more central role in melanophore development in medaka"].
- [PMID:40123122 "Only in zebrafish has one Mitf locus, mitfb, undergone a substantial, though not complete (Figure 5b) reduction in pigmentary function"].

5. Fate classification

PARTITION (subfunctionalization), at the expression / cis-regulatory level, with some
redundancy left and lineage-specific resolution.

• Established:
• The expression domains of the single tetrapod gene are split between the copies:
  melanoblasts go to mitfa only, and each copy's N-terminus matches a different tetrapod
  isoform class. Altschmied et al. framed this as subfunctionalization:
  [PMID:12019239 "In the teleost lineage subfunctionalization of mitf genes after duplication at least 100 million years ago is associated with the degeneration of alternative exons and, probably, regulatory elements and promoters."]
• Protein function is conserved (cross-rescue), so the split is not in the protein's
  core activity.
• The copies are only partly redundant: in xanthophores (two labs) and, with tfec, in the
  RPE.
• Inferred:
• That the ancestral pre-TGD gene had the union of the two expression patterns. This is
  the authors' own reading, [PMID:11543618 "recapitulate the expression and functions of a single ancestral Mitf gene, and"],
  but no gar expression data were found to test it.
• That the mitfb-only domains (epiphysis, olfactory bulb) are innovations. The authors
  only suggest [PMID:11543618 "that mitfb may serve additional novel functions."]
  No function has been shown there, so INNOVATION is not supported.
• Why not BACKUP: mitfb cannot back up mitfa in melanophores in vivo, because it is not
  expressed there (nacre has no melanophores). Redundancy is limited to xanthophores and the
  RPE.
• Why not DOSAGE: mitfa is haplosufficient in zebrafish
  [PMID:40123122 "The deletion allele also lacked discernible phenotypes when heterozygous"],
  and nothing suggests dosage balance with mitfb.
• Lineage specificity: in medaka, mitfb is the main melanophore copy and the two copies
  are more redundant, so the same ohnolog pair was resolved differently in the two lineages.
  This fits the PANTHER TGD_likely_parallel gene-tree pattern.
• Confidence: high for the melanoblast expression partition and for conserved protein
  function; moderate for "subfunctionalization" in the strict DDC sense, which needs the
  outgroup.
• What would change it:
• Gar mitf expression lacking the epiphysis and olfactory bulb domains would shift mitfb
  toward MIXED (partition plus expression-level innovation).
• A mitfb-specific function in the pineal gland would support innovation.
• A reciprocal rescue failure (mitfa unable to replace mitfb in the RPE or xanthophores)
  would add protein-level partition.

6. GO annotation consistency across the pair

From annotation-comparison.md:

• Shared and consistent (protein level conserved):
• MF GO:0000981 (IBA + IDA on both) and GO:0000978 (IBA on both) are accepted on both
  copies. GO:0046983 (IEA) is kept as non-core on both.
• CC nucleus/nucleoplasm is accepted on both; cytoplasm (IEA) is kept as non-core on both.
• The IDA GO:0006351 "DNA-templated transcription" from PMID:11543618 is modified to
  GO:0006357 on both. This keeps the pair consistent and matches the TF molecular function.
• Justified asymmetry, melanocyte differentiation (GO:0030318):
• It is core for mitfa (IBA, IMP, IGI, all accepted).
• It is kept as non-core for mitfb. The IBA (node PTN000000863) correctly records the
  inherited ancestral function, which the mitfb protein keeps, and which is central in
  medaka mitfb.
• The mitfb IGI from PMID:11543618 records a cross-rescue. That shows the protein's
  capacity, not an endogenous zebrafish role: mitfb is absent from melanoblasts and its
  nulls have normal melanophores.
• This is the textbook case where IBA propagation and a "both copies equal" reading
  over-state the role of one copy. The asymmetry here comes from expression, not from the
  protein.
• Asymmetries due to which copy was studied:
• mitfa only: pigmentation (GO:0043473), positive regulation of melanocyte differentiation
  (GO:0045636), closure of optic fissure (GO:0061386, IGI with tfec) and circadian rhythm
  (GO:0007623, UNDECIDED).
• These reflect the far larger mitfa literature. For optic fissure closure, only
  mitfa;tfec was tested; mitfb;tfec has not been.
• Removed: mitfa GO:0071228 "cellular response to tumor cell" (PMID:35929478). mitfa acts
  inside the melanoma cells as the switchable driver; the term describes a cell responding to
  a tumour cell.
• Added to both (NEW): GO:0050936 xanthophore differentiation. Evidence is the
  mitfa;mitfb double-mutant enhancement (PMID:37823232, PMID:40123122), the mitfa
  dominant-negative allele, and the mitfa;tfec interaction. The evidence is genetic
  redundancy, so the term goes on both copies.
• Deliberately not added:
• RPE/eye pigmentation: only shown for the triple mutant with tfec, in an abstract-only
  paper.
• Any pineal or olfactory bulb process for mitfb: expression only, no function tested.

7. Open questions

• Does spotted gar mitf express in the epiphysis and olfactory bulb? This decides whether
  mitfb's unique domains are ancestral (partition) or new (innovation).
• Is mitfa protein able to replace mitfb in mitfb-specific tissues? Is the N-terminal
  isoform region functionally different?
• Does an RNA-less mitfb allele reveal a single-mutant phenotype hidden by transcriptional
  adaptation in the PTC alleles?
• What does mitfb do in the pineal gland and olfactory bulb?
• Is the medaka/zebrafish difference in which copy drives melanophores due to cis-regulatory
  divergence of the two loci after the TGD (reciprocal resolution)?

References

• PMID:10433906 Lister et al. 1999 (nacre = mitfa)
• PMID:11543618 Lister et al. 2001 (discovery of mitfb; expression partition; cross-rescue)
• PMID:12019239 Altschmied et al. 2002 (teleost-wide duplicates; isoform/exon degeneration)
• PMID:20333216 Braasch et al. 2009 (FSGD pigmentation genes; synteny)
• PMID:21146516 Johnson et al. 2011 (mitfa allelic series)
• PMID:23139843 Lane & Lister 2012 (mitfb alleles; double mutants; RPE)
• PMID:23781143, PMID:24118994 Li et al. 2013/2014 (medaka mitf1/mitf2)
• PMID:30944477 transcriptional adaptation
• PMID:32541011 mitfa;tfec and choroid fissure closure
• PMID:37823232 Miyadai et al. 2023 (medaka and zebrafish double mutants)
• PMID:40123122 Korzeniwsky et al. 2025 (mitfb enhancement in xanthophores; RNA-less mitfa)
• PMID:42663426 Lister et al. 2026 (mitfa/mitfb/tfec triple mutants; RPE)
• annotation-comparison.md;
  projects/DANRE_DUPLICATION/panther_tgd_pairs.tsv