Using Xenopus to discover new candidate genes involved in BOR and other congenital hearing loss syndromes

Abstract

Hearing in infants is essential for brain development, acquisition of verbal language skills and development of social interactions. Therefore, it is important to diagnose hearing loss soon after birth so that interventions can be provided as early as possible. Most newborns in the US are screened for hearing deficits and commercially available Next Generation Sequencing hearing loss panels often can identify the causative gene, which may also identify congenital defects in other organs. One of the most prevalent autosomal dominant congenital hearing loss syndrome is Branchio-oto-renal syndrome (BOR), which also presents with defects in craniofacial structures and the kidney. Currently, mutations in three genes, SIX1, SIX5 and EYA1, are known to be causative in about half of the BOR patients that have been tested. To uncover new candidate genes that could be added to congenital hearing loss genetic screens, we have combined the power of Drosophila mutants and protein biochemical assays with the embryological advantages of Xenopus, a key aquatic animal model with a high level of genomic similarity to human, to identify potential Six1 transcriptional targets and interacting proteins that play a role during otic development. We review our transcriptomic, yeast 2-hybrid and proteomic approaches that have revealed a large number of new candidates. We also discuss how we have begun to identify how Six1 and co-factors interact to direct developmental events necessary for normal otic development.

Graphical Abstract

I. Introduction

Animal models play many important roles in understanding human health, including discovery of the genes that regulate normal human developmental processes and their defects that lead to congenital malformations and disease states. The two fully aquatic clawed frogs, Xenopus laevis and Xenopus tropicalis, have been used extensively since the 1970s to uncover the molecules that control vertebrate embryonic development (De Robertis and Gurdon, 2021; Wallingford, 2022). Adults readily reproduce in the laboratory year-round and can produce hundreds to thousands of embryos at one spawning. The cells of cleavage stages embryos have been fate mapped, are easy to microinject, transplant, and grow in explant culture. Further, the Xenopus genome has a high degree of sequence similarity and synteny with that of humans and can be genetically modified by a number of techniques (Blitz et al., 2013; Hellsten et al., 2010; Kondo and Taira, 2022; Nakayama et al., 2013; Nakayama et al., 2014; Uno et al., 2013). All these attributes make Xenopus ideal for modeling human developmental defects and congenital syndromes. In this review, we discuss how we have taken advantage of the experimental resources of Xenopus to reveal new genes that may have roles in human congenital hearing loss syndromes.

Hearing in infants is essential for postnatal brain development, acquisition of verbal language skills and development of social interactions. Therefore, it is important to diagnose hearing loss within the first month of life so that interventions, such as cochlear implants, otologic surgery, and speech therapy, can be provided as early as possible (Belcher et al., 2021). Currently, nearly all infants born in hospitals and birthing centers in the United States are tested for hearing impairments before being discharged. They typically undergo non-invasive procedures such as an otoacoustic emission test, which records responses from the cochlea after a sound is delivered by an earphone, or an auditory brainstem response test, which detects brain wave responses to a sound via surface electrodes on the infant’s head. About 1-3 per 1000 babies display some level of hearing loss at birth. The causes can include prematurity and a variety of infectious agents, but up to 60% are due to genetic factors (Belcher et al., 2021). While 70% of genetic causes are non-syndromic, the remaining 30% are represented by over 400 congenital syndromes that include defects in non-auditory organs (Korver et al., 2017). It is considered clinically important to identify causative genes underlying syndromic hearing loss so that the other affected organs can be monitored for dysfunction as the child grows. Currently, commercially available Next Generation Sequencing (NGS) panels (e.g., OtoSeq; Fulgent; Otogenetics; OtoSCOPE; Emory; OtoGenome) are used to screen patient samples for potential causative genes (Thorpe and Smith, 2020). Rather than sequencing the whole patient genome, these panels target a large number - between ~20 to >240 genes - of the most common hearing loss genes (Belcher et al., 2021), including those that are known to underlie Branchio-oto-renal Syndrome.

II. Branchio-otic and Branchio-oto-renal Syndromes

Branchio-otic Syndrome (BOS) and Branchio-oto-renal Syndrome (BOR) together comprise the second most common type of autosomal dominant syndromic hearing loss (Smith, 2018). Affected BOS and BOR individuals are characterized by a variable combination of malformations of structures associated with the: 1) second pharyngeal arch, including fistulas and cysts in the hyoid region, external ear deformities and disruption of the external auditory canal; 2) middle ear, including ossicle deformities and defects in the middle ear space; and 3) inner ear, including hypoplasia of the cochlea and/or semicircular canals, and enlarged cochlear and vestibular ducts. BOR is diagnosed when kidney malformations are additionally detected. The extent of each defect is highly variable even within a family harboring the same mutation, and hearing loss can be caused by conductive (external and/or middle ear) or sensorineural (inner ear) malformations, or a combination of both. While many of the hyoid and external ear abnormalities can be surgically corrected, patients need interventions for their hearing impairments and close monitoring of potential renal dysfunction.

Mutations in three genes have been identified as causative in about half of BOS/BOR patients (Smith, 2018). Patients diagnosed as BOS1 (Online Mendelian Inheritance in Man [OMIM] #602588) or BOR1 (OMIM #113650) carry either nonsense mutations, missense mutations, deletions, rearrangements or copy number variations in the gene encoding EYA1, a transcriptional co-factor protein known to function as an activator. It is estimated that about 40% of patients carry mutations in EYA1 (Abdelhak et al., 1998; Sanchez-Valle et al., 2010; Smith, 2018); to our knowledge, mutations have not been associated with other members of the EYA family (EYA2, EYA3, EYA4). Patients diagnosed as BOS3 (OMIM #608389) carry missense mutations or single nucleotide deletions in the gene encoding SIX1, a homeodomain transcription factor that has been shown to transcriptionally activate targets when bound to EYA proteins (Ikeda et al., 2002; Li et al., 2003; Silver et al., 2003). Missense mutations in SIX1 also have been reported for one type of autosomal dominant, non-syndromic hearing loss (Deafness, Autosomal Dominant 23; OMIM #605192). A small number of BOR2 (OMIM #610896) patients (5 individuals) have been identified that carry missense mutations in the SIX5 gene (Hoskins et al., 2007), although one of the patients also carried a large deletion in EYA1 that may have been causative (Krug et al., 2011). Mutations in these two SIX family genes are estimated to account for about 5% of BOS/BOR cases (Smith, 2018). In addition, locus 1q31 was defined as responsible for BOS2 (OMIM #129592), but the causative gene has yet to be identified (Kumar et al., 2000). Thus, in only about 50% of BOS/BOR patients tested is the underlying genetic abnormality identified. Because in newborns the variable hearing loss associated with BOS/BOR can be difficult to detect by the functional tests described above, and patients potentially carry life threatening kidney defects, genetic screening is deemed the best practice to identify at-risk BOS/BOR children (Smith, 2018). While SIX1, SIX5 and EYA1 are routinely included in the hearing loss genetic screens listed above, there is an obvious need to discover the additional genetic causes of BOS/BOR.

III. Why use Xenopus?

Xenopus, a fully aquatic frog, has been an important vertebrate model for discovering new genes that control developmental processes and for elucidating the cellular functions of the proteins they encode (Gilchrist, 2012; Harland and Grainger, 2011; Khokha, 2012; Wheeler and Brändli, 2009; see also http://www.xenbase.org). A female can lay hundreds to a few thousand eggs at one time, and the very large embryos (~1400μm in diameter for Xenopus laevis compared to ~700μm for zebrafish and ~100μm for mouse) develop external to the mother. These attributes mean that it is relatively easy to collect enough material at specific developmental times or from specific regions of the embryo to perform high-throughput, in vivo biochemical, transcriptional and proteomic analyses of gene regulation and protein function. In addition, embryonic manipulations, such as single cell or tissue isolation, culture, and transplantation, are relatively easy to perform. The individual cells of the early embryo (through 32-cells) can be identified, and their fates have been mapped by lineage tracing techniques (Hirose and Jacobson, 1979; Jacobson and Hirose, 1981; Jacobson, 1983; Dale and Slack, 1987; Moody, 1987a; Moody, 1987b; Moody and Kline, 1990). These maps can be used to target manipulations to the progenitors of specific tissues, such as the neural crest that give rise to middle and external ear structures, the otic placode that gives rise all elements of the inner ear and the intermediate mesoderm that becomes the kidney. These identified progenitors can be microinjected with a variety of molecules, including mRNAs for gain-of-function analysis of the encoded protein or its expression in an ectopic location. Microinjection of mRNAs that encode dominant-negative proteins or antisense morpholino oligonucleotides (MOs) that block RNA splicing or protein translation can be used for loss-of-function experiments in specific lineages. Most importantly, mRNAs encoding human disease variants, particularly of autosomal dominant disorders, can be microinjected to determine the impact of their defective proteins on developmental processes. In addition, transgenesis and genome editing techniques for testing gene regulation are well worked out for both Xenopus species (Horb et al., 2022). Thus, Xenopus is a versatile and powerful aquatic model in which one can rapidly test the function of genes and proteins (wild-type or mutant) in an in vivo system.

Xenopus is a tetrapod and thus shares a close evolutionary history with mammals. The Xenopus and human genomes are highly syntenic (Hellsten et al., 2010; Uno et al., 2013; Kondo and Taira, 2022) and many proteins, including Six1 and Eya1, are highly conserved with their human homologues. Transcriptome comparisons show that the mature inner ear of human and Xenopus express many of the same genes (Powers et al., 2012). In addition, the otic anatomy of Xenopus evolved for land-based hearing. While frogs do not have external ears, the mechanics of sound transmission is very similar to that in humans (Mason et al., 2009; Van Dijk et al., 2011). They detect sound via deflections of a tympanic membrane, and these are transmitted through an air-filled middle ear by the columella, which corresponds to the mammalian stapes, to a round window and on to the inner ear sensory epithelium. The tympanic membrane and middle ear ossicle are derived from the cranial neural crest (Sandell, 2014) and the labyrinthine inner ear from the otic placode. The amphibian inner ear is comprised of five vestibular end-organs, two auditory end organs and one acoustico-vestibular sacculus, each of which functions similarly to the homologous mammalian inner ear structures. In fact, frog auditory organs detect frequencies in the same range as the mammalian cochlea (Elepfandt et al., 2000; Schoffelen et al., 2008; Van Dijk et al., 2011). Thus, novel genes associated with Xenopus Six1 are likely to be highly relevant to human inner ear and middle ear development and related hearing loss syndromes.

IV. The SIX family of transcription factors and BOS/BOR variants

There are 6 members of the vertebrate Six family of transcription factors – Six1-Six6 - that are related to three Drosophila transcription factors (Cheyette et al., 1994; Blanco et al., 2010a; Hayashi et al., 2008; Pauli et al., 2005; Pignoni et al., 1997; Piñeiro et al., 2014; Serikaku and O'Tousa, 1994; Yan et al., 2003; Zhang et al., 2006; Zhou et al., 2014). Each of these proteins contain a conserved Six-type homeodomain (HD) that binds DNA and an N-terminal protein-protein interaction domain, called the Six domain (SD) (Kawakami et al., 2000; Kobayashi et al., 2001; Pignoni et al., 1997). The Six genes in fly and vertebrates are highly conserved at the amino acid level and have been grouped into three subfamilies: fly Sine oculis (SO) with vertebrate Six1/Six2; fly Six4 with vertebrate Six4/Six5; and fly Optix with vertebrate Six3/Six6 (Kawakami et al., 2000; Seo et al., 1999).

Although SIX5 variants have been found in a small number of BOR patients, it is not expressed in otic tissues. Transcripts have been detected in the retina, lens, a variety of skeletal muscles, heart and brain, and Six5-null mice develop cataracts and cardiac conduction system abnormalities (Klesert et al., 2000; Sarkar et al., 2000). For these reasons, and its location downstream of the (CTG)n repeat in the DMPK gene, SIX5 has been associated with myotonic dystrophy (OMIM #160900; Boucher et al., 1995), a muscle weakness and wasting disease. In contrast, Six1, Six2 and Six4 have overlapping expression patterns during the development of two of the organs affected in BOS/BOR: the otic placode and the nephric mesoderm (Ozaki et al., 2001; reviewed by Brugmann and Moody, 2005; Saint-Jeannet and Moody, 2014). Thus, they are candidates of interest for involvement in BOS/BOR. However, neither SIX2 nor SIX4 has been associated in OMIM with any human hearing loss syndrome. Loss of Six2 in mice results in severe kidney defects, but there are no detectable otic defects or hearing loss in either Six2-null or Six4-null mice (Ozaki et al., 2001; Self et al., 2006; He et al., 2010). These authors predict that lack of an otic phenotype is due to compensation by normal levels of Six1 and Six2 expression. In contrast, several studies in frog, fish, chicken and mouse demonstrate that Six1 plays a central role in cranial placode development. Six1 loss-of-function in Xenopus and chick results in reduced expression of several placode genes and defects in otic development (Brugmann et al., 2004; Christophorou et al., 2009; Schlosser et al., 2008). In zebrafish, Six1 knock-down results in loss of inner ear hair cells (Bricaud and Collazo, 2006; Bricaud and Collazo, 2011). Six1-null mice show defects in cranial placodes, cranial sensory ganglia and the otic vesicle (Chen et al., 2009; Ikeda et al., 2007; Ikeda et al., 2010; Konishi et al., 2006; Laclef et al., 2003; Ozaki et al., 2004; Zheng et al., 2003; Zou et al., 2004), whereas Six1-heterozygous mice have hearing loss due to cochlear defects (Zheng et al., 2003).

The BOS/BOR-causing variants in SIX1, SIX5 and EYA1 have been mapped in several patients. A large number of the mutations in EYA1 occur in the C-terminal Eya domain (ED) that is required for its phosphatase activity and binding to SIX proteins (Abdelhak et al., 1997; Jung et al., 2009). There are many missense mutations that results in amino acid substitutions in the SD of SIX1 that could interfere with co-factor binding, as well as missense and deletion mutations in the HD that could interfere with DNA binding (Fig. 1A) (Ito et al., 2006; Kochhar et al., 2008; Noguchi et al., 2011; Ruf et al., 2004; Sanggaard et al., 2007; Lee et al., 2023). A few of these mutations have been functionally analyzed in cultured mammalian cells (Table 1). The V17E mutation abolishes the SIX1-EYA interaction and prevents the translocation of EYA into the nucleus. R110W decreases the EYA interaction, W122R is predicted to reduce DNA interactions and Y129C significantly reduces DNA binding (Ruf et al., 2004; Sanggaard et al., 2007; Kochhar et al., 2008; Patrick et al., 2009; Patrick et al., 2013; Shah et al, 2020; reviewed in Moody et al., 2015). When each of these four mutations was introduced into Xenopus proteins, all four led to transcriptional deficiencies of a Six1-luciferase reporter construct in HEK cells (Shah et al., 2020). Work in zebrafish showed that expression of R110W interferes with the Six1-Eya1 interaction, resulting in defects in hair cell formation (Bricaud and Collazo, 2011). Expressing each of the four mutant proteins into wild type Xenopus embryos showed specific disruptions in the size of the domains of both neural crest and cranial placode genes at neural plate stages, as well as reduced gene expression in the otic vesicle (Shah et al., 2020; Mehdizadeh et al., 2021). Structural defects in the tadpole inner ear, including the size of the otic capsule, otoconia and sensory hair cell patches, also were observed. Similarly, an ENU-induced mouse mutant line called Catweasel (Cwe) that harbors an HD missense mutation (Fig. 1A; Bosman et al., 2009) displayed hair cell defects. A mutation similar to Cwe was reported in one family that presented with only auditory defects (Mosrati et al., 2011). Four missense amino acid mutations have been identified in SIX5 (Fig. 1B): A158T is located in the SD, whereas A296T, G365R and T552M are located downstream of the HD and upstream of the activation domain (Hoskins et al., 2007). Yeast two-hybrid (Y2H) liquid α-galactosidase assays that used Eya1 as prey indicated a >2-fold reduction in expression with A296T and T552M, suggesting that the C-terminal region of Six5 also may interact with Eya1.

Figure 1:

A. The amino acid sequence of the human and Xenopus Six domain (SD, in dark blue letters) and N-terminal residues of the homeodomain (HD, in red letters) are identical. The known BOR-causing human variants are noted in teal using the human SIX1 amino acid numbers; seven are in the SD and three are in the HD. The Cwe variant is marked in magenta. The bait used for the Hybrigenics Y2H screen included the entire SD plus the entire HD extending to amino acid 190.

B. Schematic of the SIX5 protein showing the Six domain (SD), homeodomain (HD), and C-terminal activation domain. The relative positions of the mutations found in BOR patients (Hoskins et al., 2007) are indicated.

Table 1: Summary of SIX1 pathogenic mutants.

Several methods have been used in the literature to determine the deficits resulting from human BOR variants using cultured cell lines. wt, wild type protein; SD, variants in the Six Domain, HD, variants in the Homeodomain, ED, Eya domain, Y2H, yeast two-hybrid; NC, no change; NI, no induction (baseline or lower); WB, Western blot; Gal4^BD, DNA binding domain of Gal4; Gal4^AD, Activation domain of Gal4; ↓< wt SIX1; ↑, = > wt SIX1.

		wt	V17E	H73P	V106G	R110Q	R110W	R112C	W122R	Y129C	delE133
Mutation location:		---	HD	HD	HD	HD	HD	HD	SD	SD	SD
Patrick et al., 2009	GST-fusion Six1 protein expression in E. coli	Yes	Yes	Failed	Yes	Failed	Yes	Yes		Yes	Yes
Patrick et al., 2009	CD spectroscopy: α-helix content		NC		NC		NC		↓	NC	↓
Patrick et al., 2009	CD spectroscopy: Thermal denaturation (Tm)	Tm = 34.5°C	NC		↓ 28.2°C		NC	↓ 26.8°C		NC	NC
Patrick et al., 2009	PPI – SIX1-EYA2ED complex peak (by S200 exclusion gel filtration chromatography)	Yes	No		Yes		Yes	Yes (↓ ?)		Yes	Yes
Ruf et al., 2004	PPI-SIX1-EYA2 (Y2H LacZ activity with Gal4BD-SIX1 & Gal4AD-EYA1ED)	Yes					~ 4-fold less			~ 6-fold less	~ 8-fold less
Patrick et al., 2009	Nuclear localization of full-length EYA2 in MCF7 cells (by nuclear/cytoplasm fractionation and WB	Yes	No	Yes	Yes	Yes	Yes	Yes		Yes	Yes
Patrick et al., 2009	Nuclear localization of full-length EYA2-Flag in MCF7 cells (by nuclear/cytoplasm fractionation and WB	Yes	No	Yes	Yes	Yes	Yes	Yes		Yes	Yes
Shah et al., 2020	SIX1 + EYA1 co-localization in HEK cells	Nuclear Eya1	Nuclear and some cytoplasmic EYA1				Nuclear Eya1		Nuclear Eya1	Nuclear Eya1
Patrick et al., 2009	Increased levels of SIX1 and EYA2ED proteins in MCF7 cells	Yes	No		Yes		Yes	Yes		Yes	Yes
Patrick et al., 2009	SIX1 binding to MEF3-DNA by EMSA	Low	↑		↓		↓	↓		↓	↓
Ruf et al., 2004	MEF3-DNA binding of SIX1 by EMSA	Yes					NC			Reduced	Abolished
Patrick et al., 2009	SIX1+EYA2ED binding to MEF3-DNA by EMSA	Higher	NC
Patrick et al., 2009	SIX1+EYA2 transcription (luciferase in MCF7 cells)	10-fold induced	NI	NI	NI	NI	NI	NI		NI	NI
Ruf et al., 2004	SIX1+EYA2 transcription (luciferase in HEK cells)	~ 3-fold increase					NI			NI	NI
Shah et al., 2020	SIX1+EYA1 transcription (luciferase in HEK cells)	~ 7-fold increase	NI				NI		NI	NI
Patrick et al., 2009	DNA binding affinity of SIX1 ± EYA2ED (Kd)	SIX1: 2404±246 SIX1+EYA2ED: 383±64	V17E: 211±44 V17E+EYA2ED: 336±43

V. In search of novel candidate genes underlying BOS/BOR

Since nearly half of BOS/BOR patients do not harbor mutations in SIX1, SIX5 or EYA1, we decided to search for novel genes using Xenopus embryos, which have been used for decades to discover developmentally important genes. We reasoned that novel causative genes would likely be found by identifying factors that are genetically and functionally related to Six1 during otic development. As reviewed below, one approach we have taken is to identify genes that are regulated by Six1 during early craniofacial development, and therefore are part of its inner and middle ear gene regulatory network. A second approach we have taken is to identify additional proteins that interact with Six1, perhaps acting as co-factors that can modify Six1 transcriptional activity. We posit that genes discovered by either approach would become high priority candidates for functional characterization in wild type embryos, and if later verified in human patients, for inclusion in hearing loss screening.

V. A. In search of novel Six1 transcriptional targets

Many studies in vertebrate animal models have demonstrated that Six1 and Eya1 play key roles in the development of the cranial placodes, neural crest and kidney (Brugmann et al., 2004; Brodbeck and Englert, 2004; Christophorou et al., 2009; Grocott et al., 2012; Moody and LaMantia, 2015; Ohto et al., 1998; Ozaki et al., 2004; Pandur and Moody, 2000; Schlosser, 2021; Xu et al., 1999; Xu et al., 2003). Therefore, genes that act downstream of Six1 are high priority candidates for novel BOS/BOR-causing mutations because they carry out the Six1 developmental program.

Candidate Six1 target genes have already been identified in the fly, in which SO plays a critical role in eye development. A ChIPseq analysis using DNA isolated from developing fly eye-antennal imaginal discs identified nearly 6,000 putative SO target genes, over half of which had not previously been described to play a role in eye development (Jusiak et al., 2014a; Jusiak et al., 2014b). The vertebrate homologues of a subset of these genes have been shown to be involved in cranial placode development, including atonal (vertebrate Atoh1), dachshund (vertebrate Dach1), eyeless (vertebrate Pax6), hedgehog (vertebrate Shh) and prospero (vertebrate Prox1 and Prox2) (reviewed in Saint-Jeannet and Moody, 2014). Although many of these are expressed in the embryonic tissues related to BOS/BOR defects (cranial placodes, neural crest, kidney), so far only ATOH1 has been reported to harbor mutations that result in hearing loss (Table 2). Since nearly half of the genes in the Drosophila data set had not previously been associated with eye development, we believe it will be worth mining this data set for additional vertebrate homologues that might be involved in congenital hearing loss syndromes.

Table 2: Genes that are potential targets of Six1.

Vertebrate homologues of genes identified by Drosophila SO ChIPseq and Xenopus microarray and RNAseq studies are listed. Also included are genes affiliated with Six1 during cranial placode development found in the literature (see text). The embryonic tissues in which they are expressed and the OMIM number for an associated congenital syndrome are provided where there are published data.

Gene name	Protein type and function	Embryonic tissue distribution	Associated congenital syndrome (OMIM number)
Atoh1	bHLH transcription factor; brain and placode neurogenesis	Brain, otic vesicle, trigeminal placode and ganglion	Deafness, autosomal dominant 89 (620284)
Dach1	Transcriptional repressor; binds Eya1; interacts with Six6	Brain, retina, hyoid arch, limb bud, genital eminence, kidney	None known to date
FoxI1	Winged helix transcription factor, transcriptional activator	Otic vesicle, kidney, epidermis	Enlarged vestibular aqueduct (EVA) (600791)
FoxG1	Winged helix transcription factor, both activator and repressor; neuronal differentiation	Brain, retina, cranial placodes, otic vesicle, pharyngeal neural crest,	Rett syndrome, congenital variant (613454)
Irx1	Iroquois homeobox transcription factor; neurogenesis	Brain, eye, otic vesicle, pharyngeal ectoderm, heart, limbs, kidney	None known to date
Pax2	Paired box transcription factor; cell fate decisions	Brain, preplacodal ectoderm, optic vesicle, otic vesicle, hindbrain, spinal cord, adrenals, kidney	Glomerulosclerosis, focal segmental, 7 (616002) Papillorenal syndrome (120330)
Pax6	Paired box transcription factor; cell fate decisions	Retina, brain, spinal cord, lens placode	Multiple eye defects (120200; 120430; 136520; 148190; 106210; 165550; 604229)
Pax8	Paired box transcription factor; cell fate decisions	Preplacodal ectoderm, otic vesicle, kidney, thymus	Hypothyroidism, congenital (218700)
Prox1	Prospero-related homeobox transcription factor	Brain, lens, liver, kidney, pancreas, heart, lymphatics	None known to date
Prox2	Prospero-related homeobox transcription factor, transcriptional activator	Cranial ganglia, retina	None known to date
Shh	Secreted factor involved in many cell fate decisions	Notochord, floorplate of neural tube, gut endoderm, posterior limb buds	Holoprosencephaly 3 (142945); Microphthalmia with coloboma 5 (611638); Schizencephaly (269160); Single median maxillary central incisor (147250)
Sox2	SRY-box transcription factor; neurogenesis, maintenance of stem cells	Neural plate, retina, cranial ganglia placodes, otic vesicle	Microphthalmia, syndromic 3 (206900)
Tbx1	T-box transcription factor;	Preplacodal ectoderm, pharyngeal arches, pharyngeal pouches; otic vesicle, epibranchial placodes, kidney, vertebrae, tooth buds	Conotruncal anomaly face syndrome (217095); DiGeorge syndrome (188400); Tetralogy of Fallot (187500); Velocardiofacial syndrome (192430)
Tbx5	T-box transcription factor;	Retina, lens, heart, forelimb, genital papilla, lung, pharynx, thoracic body wall	Holt-Oram syndrome (142900)
XB5850668	Sulfotransferase; uncharacterized	Neural tube, retina, cranial placodes, neural crest, pharyngeal arches, otic vesicle, kidney	None known to date

Many studies in vertebrates have shown that Pax and Fox genes work in concert with Six1 and Eya1 during placode and kidney development (reviewed in Bhattacharyya and Bronner-Fraser, 2004; Brodbeck and Englert, 2004; Grocott et al., 2012; Moody and LaMantia, 2015; Moody et al., 2015). However, mutations in members of these gene families do not phenocopy BOS/BOR (Table 2). PAX2 mutations present primarily with eye and renal defects, although a few patients have sensorineural deafness (reviewed in Grocott et al., 2012); in mice, Pax2 interacts with Eya1 during otic development (Zou et al., 2006). Patients with PAX6 mutations present primarily with ocular defects and those with PAX8 mutations present primarily with hypothyroidism and kidney defects. Patients with FOXI1 mutations present with deafness and vestibular defects, but no external ear or kidney defects, and patients with FOXG1 mutations present with Rett Syndrome and severe intellectual disability. Other studies have experimentally placed Six1 upstream of Sox2, Irx1, Tbx1 and Tbx5 during cranial placode development (Brugmann et al., 2004; Schlosser and Ahrens, 2004; Schlosser et al., 2008; Sullivan et al., 2019; reviewed in Grocott et al., 2012; Streit, 2004). However, in OMIM mutations in these genes are associated with tissues other than inner ear and kidney (Table 2). Thus, a search for additional transcriptional targets of Six1 is warranted.

Two such screens have been performed with Xenopus tissues with the similar goal of identifying novel Six1 targets. We performed a microarray assay of Xenopus ectodermal explants that expressed high levels of Six1 (Yan et al., 2015). This approach identified 72 significantly up-regulated genes and 58 significantly down-regulated genes. As in the fly ChIP-Seq study, most candidates were of unknown function, but over 30 were expressed in the otic placode and developing kidney. To date, we have shown that one candidate, Irx1, plays a reciprocal regulatory role with Six1 during preplacodal ectoderm (PPE) and otic placode development (Sullivan et al., 2019), but as noted above, IRX1 mutations in humans do not phenocopy BOS/BOR (Table 2). A surprising candidate Six1 target is XB5850668, an uncharacterized sulfotransferase. We recently found that it is required for the proper proportioning of the embryonic ectoderm into neural plate, neural crest, PPE and epidermis (Marchak et al., 2023). Significantly, loss of XB5850668 reduced the expression of a number of otic vesicle genes, including Irx1. It is unclear which, if any, of the many human sulfotransferases are homologous to XB5850668, but a review of SULT genes in OMIM does not identify any that have hearing loss or craniofacial defects. Riddiford and Schlosser (2016) likewise screened for potential Six1-Eya1 targets by expressing hormone-inducible versions of these genes in PPE explants in the presence of protein synthesis inhibitors, which is a reliable method to assess putative direct targets that is easily performed in Xenopus embryos (Kolm and Sive, 1995; Seo et al., 2007). Comparing the transcriptomes by RNAseq of hormone-induced versus control explants identified hundreds of potential targets, including proteins involved in transcription, protein-protein binding, catalytic processes and signaling. Among these were Atoh1, Sobp, a few sulfotransferases, Sox2 and members of the Notch pathway.

In summary, these studies in humans and many experimental model systems show that Six1 is part of a complicated gene regulatory network that involves hundreds if not thousands of downstream genes. While the Xenopus studies have greatly expanded the list of potential Six1 targets, ChIPseq approaches will be needed to validate which genes are directly regulated by Six1. Sorting through all these excellent data sets for those genes that are direct targets of Six1 and then associating those with congenital hearing loss and BOS/BOR remains a very large task, but one with a very worthwhile goal.

V.B. In search of novel Six1 co-factors

An important aspect of gene expression regulation by Six factors involves their association with co-factor proteins that do not bind DNA themselves but influence the transcriptional activity of the complex (reviewed in Saint-Jeannet and Moody, 2014; Moody et al., 2015). Experiments with fly and vertebrate SO/Six proteins show that an association with Eya co-factors promotes transcriptional activation, whereas association with Groucho/Tle co-factors results in transcriptional repression (Ikeda et al., 2002; Li et al., 2003; Silver et al., 2003; Zhu et al., 2002; Brugmann et al., 2004; Christophorou et al., 2009). Interestingly, in the zebrafish otic vesicle, combined Six1/Eya1 activation of target genes is required for hair cell formation, whereas combined Six1/Tle repression of target genes is required for auditory-vestibular ganglion neuron formation (Bricaud and Collazo, 2011).

The focus of SO/Six co-factor research has been primarily on the Eya and Groucho/Tle families because they are known to bind to Six proteins, affect their transcriptional activity and are expressed in the same tissues, the ear and kidney being of particular relevance to BOS/BOR (Bajoghli et al., 2005; Bane et al., 2005; Brugmann and Moody, 2005; Brugmann et al., 2004; Kobayashi et al., 2001; Li et al., 2010; Neilson et al., 2010; Ohto et al., 1999; Ozaki et al., 2002; Zhu et al., 2002). Nonetheless, other proteins have been shown to interact with SO: Y2H and GST-pulldown assays in fly identified more than 25 interacting proteins (Anderson et al., 2014; Giot et al., 2003; Kenyon et al., 2005a; Pignoni et al., 1997). As expected, Eya and Groucho had the highest interaction scores, but these assays strongly indicated an abundance of potential SO co-factors. Based on the fly interactome and the fact that mutations in the EYA1 co-factor account for the largest percentage of known genetic causes of BOS/BOR, we decided to take a multi-pronged approach to identify novel putative co-factors of vertebrate Six1. These efforts included: 1) screening the sequences represented in the Xenopus Gene Collection (Gerhard et al., 2004; Morin et al., 2006) for homologues of Drosophila SO-binding proteins; 2) a Y2H screen of a Xenopus embryo cDNA library with Six1 SD-HD as bait; and 3) a mass spectrometry based proteomic screen.

B1. Screening the Xenopus Gene Collection for homologues of SO-binding proteins

We took advantage of the Drosophila SO interactome dataset to screen for potential vertebrate Six1 co-factors using the enormous collection of Xenopus cDNA sequences in the Xenopus Gene Collection (Gerhard et al., 2004; Morin et al., 2006), which was available before Xenopus genomes were sequenced (X. tropicalis, Hellsten et al., 2010; X. laevis, Session et al., 2016). We predicted that because the amino acid sequence of the fly SD is highly conserved in Xenopus Six1 and the Xenopus SD amino acid sequence is identical to human, the proteins that bind to fly SO are likely to bind to Xenopus Six1 and likewise be relevant to human. First, we performed BLAST analyses to identify cDNAs in the Xenopus Gene Collection that are likely homologues of the fly co-factors, and identified 33 genes with high amino acid sequence similarity to 20 of the 25 fly SO-interactors. Next, an in situ hybridization (ISH) analysis of 20 of those genes (representing 11 of the fly interactors) demonstrated that 16 are expressed in the otic placode and otic vesicle and 11 in the nephric mesoderm (Neilson et al., 2010). For the past several years we have been characterizing some of the candidates we identified by this approach, testing each candidate for: 1) similarity in protein sequence to the human homologues; 2) expression in the developmental rudiments of the inner ear (otic placode/otic vesicle), middle ear (neural crest-derived pharyngeal arches) and/or kidney (nephric mesoderm); 3) requirement for normal ear development, as determined by knockdown approaches; 4) binding to the Six1 protein, as determined by Co-IP assays; and 5) ability to affect Six1 transcriptional activity, as determined by Six1-luciferase reporter assays. Below, we summarize what we know to date about five of the novel candidates.

Sobp (Sine oculis binding protein) was first discovered in fly to directly bind to SO and alter its transcriptional activity (Kenyon et al., 2005a, b). It contains two FCS-type zinc fingers, a proline-rich domain, and a nuclear localization signal (NLS) that are highly conserved from fly to human (Fig. 2A). Human and Xenopus Sobp are 82.5% identical at the amino acid level. In Xenopus and mouse, Sobp is expressed in the neural tube, PPE and otic vesicle, where its expression overlaps with that of Six1 (Fig. 2C) (Chen et al., 2008; Neilson et al., 2010; Tavares et al., 2021). In mouse, two naturally occurring mutations in Sobp cause inner ear deformities (Chen et al., 2008); a mutation in a family with MRAMS syndrome (mental retardation, anterior maxillary protrusion, strabismus; OMIM #613671) caused craniofacial abnormalities in all affected individuals and hearing loss in one of them (Birk et al., 2010). Both the mouse and human mutations result in truncated proteins that are missing the C-terminal NLS (Fig. 2A). Knockdown of Sobp in wild-type Xenopus embryos similarly resulted in disrupted otic vesicle and craniofacial cartilage development (Tavares et al., 2021). We demonstrated that Six1 binds to Xenopus Sobp, but Sobp also binds to Eya1 and prevents Six1 from translocating Eya1 to the nucleus (Tavares et al., 2021). As a result, Sobp reduces the transcriptional activation of the Six1-luciferase reporter in the presence of Six1+Eya1.

Figure 2: Candidate Six1 co-factors.

A. Schematic of the domains present in human and Xenopus Sobp proteins. Sobp contains a C-terminal nuclear localization signal (NLS), two FCS-type Zinc finger domains (ZF1 and ZF2), and a proline-rich region of 245 amino acids. Box 1 includes 20 amino acids that are identical in Xenopus and human proteins. Box 2 includes one of the FCS-type ZF domains. Box 3 is a 25 amino acid domain identical in Xenopus and human proteins. The amino acid locations of the known mouse and human SOBP mutations are indicated by red arrows. Each mutation introduces a stop codon (X) that truncates the protein (Chen et al., 2008; Birk et al., 2010). Box 2 and Box 3 were used in a BLAST search that identified Zmym2 and Zmym4 (Neilson et al., 2010).

B. Schematic of the domains present in human and Xenopus Zmym4 proteins. Like Sobp, it contains several FCS-type ZF domains and a C-terminal NLS. A presumed DNA-binding domain is located N-terminal to the ZFs. The yellow box indicates a DUF3504 domain. Regions of homology to Sobp Box2 and Box3 are indicated.

C. Wholemount in situ hybridization images of wild-type embryos indicating the expression domains of transcripts encoding Sobp, Zmym2, Zmym4, 2G4 and Mcrs1 (from Neilson et al., 2010). Red arrows indicate the otic vesicle, which gives rise to the inner ear. Arrows indicate the branchial arches, which give rise to elements of the middle ear (asterisks) and cranial cartilages. e, eye.

Zmym2 and Zmym4 were identified in our screen by their sequence similarity in the Box 2 and Box 3 regions of fly Sobp (Neilson et al., 2010). Like Sobp, each contains several FCS-type zinc fingers, a proline-rich domain, and a C-terminal NLS (Fig. 2B). Xenopus Zmym2 is 71.3% identical and Zymy4 is 76.6% identical to their human homologues at the amino acid level. Zmym2 and Zmym4 are expressed in the PPE, neural crest, otic vesicle and nephric mesoderm in both Xenopus and mouse (Fig. 2C) (Gray et al., 2004; Neilson et al., 2010). Little is known about their developmental functions, however, Zmym2 is part of a transcriptional co-repressor complex that has been shown to down-regulate E-cadherin in cultured cells (Gocke and Yu, 2008). Knockdown of Zmym2 in Xenopus embryos resulted in expansion of neural plate and neural crest genes, and reduction of PPE genes, whereas Zmym4 knockdown additionally reduced neural crest genes (Jourdeuil et al., 2022). Both Zmym2 and Zmym4 are required for proper formation of the otic vesicle and their knockdown phenotypes are similar to those of Sobp (Tavares et al., 2021). However, our preliminary data indicate that neither Zmym2 nor Zmym4 bind to Six1 or alter the transcriptional activation of the Six1-luciferase reporter in the presence of Six1+Eya1 (Jourdeuil et al., 2022). Although these data suggest that their developmental activities may be independent of Six1, mutations in ZMYM2 have been associated with CAKUT syndrome (OMIM #619522); these patients do not phenocopy BOS/BOR but do present external ear and kidney defects.

Proliferation-associated 2G4 (Pa2G4) (aka Ebp1) emerged as a candidate based on its high similarity to the Drosophila SO-binding partner CG10576 (Giot et al., 2003). Structural studies indicate that Pa2G4 contains a “pita bread” fold structure that mediates interactions with other proteins to regulate cell proliferation and differentiation (Figeac et al., 2014). The C-terminal domain of Pa2G4 also contains distinct motifs involved in protein binding (Kowalinski et al., 2007, Monie et al., 2007; Squatrito et al., 2004; reviewed in Neilson et al., 2017). Like Six1, Pa2G4 is expressed in the otic placode, otic vesicle, neural crest-derived pharyngeal arches and nephric mesoderm (Fig. 2C)(Gray et al., 2004; Neilson et al., 2010). We showed that Pa2G4 interacts with Six1 and competes with its association with Eya1, thus reducing the level of transcriptional activation of the Six1-luciferase reporter in the presence of Six1+Eya1 (Neilson et al., 2017). Knockdown of Pa2G4 expanded the neural plate and reduced the size of the both the neural crest and PPE domains; the otic vesicle also was reduced in size and the expression of a number of otic genes were reduced (Neilson et al., 2017).

Mcrs1 (Microspherule-1) contains a forkhead-associated domain that mediates protein-protein interactions (Li et al., 2000). In Xenopus and mouse embryos, Mcrs1 is expressed in the otic placode, otic vesicle, neural crest-derived pharyngeal arches and nephric mesoderm (Fig. 2C)(Gray et al., 2004; Neilson et al., 2010). We showed that Mcrs1 interacts with Six1 but there was no evidence that it competes with Eya1; nonetheless, it significantly reduced the level of transcriptional activation of the Six1-luciferase reporter in the presence of Six1+Eya1 (Neilson et al., 2020). Knockdown of Mcrs1 also expanded the neural plate but had pleiotropic effects on the sizes of the neural crest and PPE domains (Neilson et al., 2020). The otic vesicle and the cranial cartilages were also dysmorphic after Mcrs1 knockdown (Neilson et al., 2020; Keer et al., 2022).

These data encourage us to continue the functional characterization of these five proteins because they are normally expressed in ear and kidney progenitor tissues, their loss affects otic development, and they are highly similar to their human homologues. However, only two of these five candidates are recorded in OMIM to underlie a congenital syndrome - SOBP and ZMYM2 – neither of which phenocopy BOS/BOR. Therefore, we have taken two other approaches to search for proteins that may interact with Six1 during Xenopus development: a yeast 2-hybrid (Y2H) screen and a mass spectrometry proteomic screen.

B2. Yeast 2-hybrid screen of a Xenopus embryo library using Six1 as bait

Since the above screen demonstrated that many of the Six1-interacting proteins are expressed during cranial placode development, the Moody laboratory performed a Y2H screen using a fragment of Xenopus Six1.L (NM_001088558.1) as the bait. The Six1 bait fragment included amino acids 1-190, which spans the well-characterized N-terminal protein-protein interaction domain (SD) and the DNA-binding homeodomain (HD) (Fig. 1A). The Y2H screen was performed by Hybrigenics Services, S.A.S., Evry, France (http://www.hybrigenics-services.com). Briefly, the coding sequence for aa1-aa190 of Xenopus laevis six1.L was PCR-amplified and cloned into pB27 as a C-terminal fusion to LexA (LexA-six1.L). The construct was checked by sequencing the entire insert and used as a bait to screen a random-primed Xenopus laevis stage 23-26 cDNA library constructed in pP6; pB27 and pP6 are derived from the original pBTM116 (Vojtek and Hollenberg, 1995) and pGADGH (Bartel and Fields, 1995) plasmids, respectively. 99 million clones (10-fold the complexity of the library) were screened using a mating approach with YHGX13 (Y187 ade2-101::loxP-kanMX-loxP, MATα) and L40deltaGal4 (MATα) yeast strains, as previously described (Fromont-Racine et al., 1997). 178 His+ colonies were selected on a medium lacking tryptophan, leucine and histidine, and supplemented with 0.5 mM 3-aminotriazole to reduce potential bait autoactivation. The prey fragments of the positive clones were amplified by PCR and sequenced at their 5’ and 3’ junctions. The resulting sequences were used to identify the corresponding interacting proteins in the GenBank database (NCBI) using a fully automated procedure. A confidence score - Predicted Biological Score (PBS) - was attributed to each interaction as previously described (Formstecher et al., 2005).

The PBS relies on two different levels of analysis. First, a local score takes into account the redundancy and independency of prey fragments, as well as the distribution of reading frames and stop codons in overlapping fragments. Second, a global score takes into account the interactions found in all the screens performed at Hybrigenics Services using the same library. This global score represents the probability of an interaction being nonspecific. For practical use, the scores were divided into four categories, from A (highest confidence) to D (lowest confidence) (Table 3). Other categories (not shown) specifically flag interactions involving highly connected prey domains previously found several times in screens performed on libraries derived from the same organism, including highly connected domains previously confirmed as false positives of the technique. The PBS scores of A-D have been shown to positively correlate with the biological significance of the interactions (Rain et al., 2001; Wojcik et al., 2002).

Table 3: Y2H predicted Six1-interacting proteins.

Xenopus proteins predicted to interact with Six1 based on Hybrigenics Y2H screen. The embryonic tissues in which they are expressed and the OMIM number for an associated congenital syndrome are provided where there are published data.

Confidence	Number of clones	In-frame (IF) or out-of- frame (OOF)	Protein name	Protein function	Gene accession number (NCBI)	Xenopus embryo tissue expression (from Xenbase)	Associated congenital syndrome (OMIM number)
Very high (A)	61	IF	Eya3.L	Known Six cofactor	NM_001096525.1	Neural plate, pre-placodal ectoderm	Craniofacial microsomia (#164210)
	4	OOF	LOC108701771	Predicted eya2, transcript variant X2; known Six cofactor	XM_041576045.1	Neural tissues, cranial placodes, somites	None to date
High (B)	11	IF	Eya1.L	Known Six cofactor	NM_001090419.1	Neural tissues, cranial placodes, somites	BOR1 (#113650); BOS (#602588); Otofaciocervical syndrome (#166780); Anterior segment anomalies (#602588)
	16	OOF	Eya2.S	Known Six cofactor	XM_018238153.2	Neural tissues, cranial placodes, somites	None to date
	4	IF	Krt7.L	Keratin 7; type II cytokeratin	NM_001101757.1	Non-neural ectoderm, cement gland, cranial ganglia, pharyngeal arch, somites	None to date
	7	IF	Ralbp1.S	RalA binding protein 1; receptor-mediated endocytosis; downstream effector of RAL, a GTP-binding protein involved in RAS signaling pathway.	XM_018267888.2	no embryonic tissue expression data available	None to date
Good (C)	2	IF	Pax3 (L or S?)	Paired domain transcription factor	NM_001095525.1; NM_001095524.1	Neural crest, brain, hypaxial muscle	Waardenburg type 1 (#193500) and type 3 (#148820); Craniofacial-deafness-hand (#122880); Rhabdomyosarcoma 2 (#268220)
	5	IF	Rai14.L	Retinoic acid induced 14, transcript variant X6; part of the 26S proteosome regulatory complex; actin binding; apoptotic signaling pathway; NIK/NF-kappaB signaling	XM_018266493.2	no embryonic tissue expression data available	None to date
Moderate (D)	1	IF	Eya3.S	Known Six cofactor	XM_041583935.1	Neural plate, pre-placodal ectoderm	None to date
	2	IF	Eya4.L	Known Six cofactor	XM_018263246.2	no embryonic tissue expression data available	Cardiomyopathy 1J (#605362); Deafness, autosomal dominant 10 (#601316)
	5	IF	Tle2.L (esg2-a)	Transducin-like enhancer of split 2; Known Six cofactor (Gro/Tle family)	NM_001088355.1	Neural plate, kidney, otic vesicle, branchial arch neural crest,	None to date
	5	IF	LOC108698840	Predicted CXXC-type zinc finger protein 1, transcript variant X5; transcriptional activator ; component of the SETD1 complex	XM_041572242.1	no embryonic tissue expression data available	None to date
	1	IF	LOC494754	Serine/arginine repetitive matrix 1 protein (SRRM1); involved in several pre-mRNA processing events	BC108752	Eye, pharyngeal arch	None to date
	4	IF	Anxa6.L	Annexin A6; calcium-dependent membrane and phospholipid binding protein; endosome aggregation and vesicle fusion	NM_001092378	no embryonic tissue expression data available	None to date
	1	IF	C1qbp.L	Complement component 1, q subcomponent binding protein; combines with C1r and C1s to form first component of the serum complement system; part of pre-mRNA splicing factor SF2; binds hyaluronic acid; Rho GTPase signaling, apoptotic signaling pathway	XM_018247011.2	no embryonic tissue expression data available	Combined oxidative phosphorylation deficiency (#617713)
	1	IF	Get1.S	Guided entry of tail-anchored proteins factor 1; receptor required for post-translational delivery of TA-proteins to the ER	NM_001093887.2	no embryonic tissue expression data available	Congenital heart defects in Down syndrome (#190685)
	1	IF	Hes4.L	Aka Hairy 2; bHLH TF, facilitates transcriptional repressor	NM_001088692.1	Neural plate, neural plate border, notochord, neural crest, pre-placodal ectoderm, somites, eye, kidney	None to date
	2	IF	Hey1 (L or S?)	Hes-related family bHLH TF with YRPW motif 1; nuclear protein induced by Notch signaling	BC084410.1 (L); NM_001090457.1 (S)	Neural plate, cranial placodes, somites, pharyngeal arch, eye	None to date
	1	IF	Hey2.L	Hes-related family bHLH TF with YRPW motif 2	XM_018263187.2	Neural plate, cranial placodes, neural crest, pharyngeal arch, somites, eye, heart, kidney	Brugada? (#601144)
	1	IF	Kcne5.L	Potassium channel, voltage gated subfamily E regulatory beta subunit 5, transcript variant X1; subunit of voltage-gated potassium channels	XM_018228726.2	no embryonic tissue expression data available	None to date
	1	IF	Kdm6b.S	Lysine (K)-specific demethylase 6B, transcript variant X9; demethylates di- or trimethylated lysine 27 of histone H3; demethylates non-histone proteins	XM_041588963.1	Brain, notochord, neural crest, pharyngeal arch, somites, eye, otic vesicle, lens	Neurodevelopmental disorder with course facies and mild skeletal abnormalities (#618505)
	1	IF	Kpnb1.S	Karyopherin (importin) beta 1, transcript variant X2; docks the NLS containing protein-importin alpha complex on the cytoplasmic side of nuclear pore	XM_041577603.1	Brain, pharyngeal arch, otic vesicle,	None to date
	1	IF	Rlip	RAL interacting protein; same as Ralbp1	AJ252165.1	no embryonic tissue expression data available	None to date
	1	IF	Usp1.L	Ubiquitin specific peptidase 1, transcript variant X1; cleaves ubiquitin moiety from ubiquitin-fused precursors and ubiquitinylated proteins	XM_018258126.2	no embryonic tissue expression data available	None to date

This screen identified 131 clones representing 25 putative Xenopus Six1 interacting proteins in the PBS A-D categories (Table 3). It includes all four members of the Eya family, and one member of the Groucho/Tle family (Tle2.L); these proteins were expected from previous studies mentioned above that demonstrated binding to Six1/SO, thus validating the screening protocol. It also is interesting that three proteins involved in transducing the Notch pathway (Hes4.L; Hey1, Hey2.L) were identified because Riddiford and Schlosser (2017) showed that Six1 and Eya1 can regulate placode neurogenesis by activating the Notch pathway genes, neurog1 and hes8. Although Hes4, Hey1 and Hey2, identified in our Y2H screen, have not yet been definitively attributed to a human congenital syndrome (Table 3), these clues spur us to dig further into the relationship between Notch and Six1 during craniofacial development. There also were a few surprises. Many of the putative interacting proteins participate in fundamental cellular processes, including endocytosis and vesicle fusion, the Ras pathway, proteosomes, apoptosis, pre-mRNA processing, Rho GTPase signaling, ubiquination, potassium channels, importing proteins to the nucleus and demethylation (Table 3). Many of these processes have been associated with adult diseases, but less than half with congenital syndromes in the OMIM database (Table 3). However, where expression data are available in Xenopus embryos, many are expressed in tissues affected in BOS/BOR: neural crest and pharyngeal arches, cranial placodes and kidney. These proteins will need to be validated by Co-IP assays as true Six1-interacting proteins, and then tested for a role in craniofacial development by knockdown strategies. As presented above for the homologues of Drosophila SO-interacting proteins, this is a very straight-forward approach in Xenopus embryos.

Interestingly, two DNA binding proteins were identified by this Y2H screen. LOC108698840 is a predicted Xenopus homologue of human CXXC1, which acts as a transcriptional activator with preferred binding to CpG unmethylated motifs (Fujita et al., 2000); it has not yet been associated with a congenital syndrome (Table 3). Pax3 is a paired box transcription factor involved in neural crest formation (Sato et al., 2005; Milet et al., 2013) and is associated with Waardenburg and Craniofacial-deafness-hand syndromes (Table 3). However, because it shows up in many similar screens performed by Hybrigenics Services, it suspected to be a false positive.

It was surprising that of the putative cofactors identified in the screen of Drosophila SO-interactors, only Eya and Groucho/Tle proteins were identified. However, there are several possible reasons for this lack of overlap between the screens. Proteins might be missing from the library, some might not be expressed well in yeast, and some might interact with the C-terminal region of Six1, as seen with SIX5 (Hoskins et al., 2007); this region was not included in the bait. Negative results also can occur if the in-frame protein is toxic to the yeast, or if the interaction domain is larger than 300 amino acids, which is the average fragment size (900bp) of the library inserts. Further, we have discovered that inclusion of the HD can occasionally interfere with co-factor binding in Y2H assays (see below). For these reasons, our Y2H screen cannot be considered comprehensive. However, it has provided numerous novel candidates to explore for their developmental roles in otic and kidney development, for which Xenopus embryos are ideal.

B3. Identifying Xenopus Six1-interacting proteins by mass spectrometry

Another method for discovering novel partners of Six1 is to analyze the proteome by mass spectrometry (MS). Because many protein interactions may be transient or weak, the Alfandari laboratory used two approaches to enrich for bona fide Six1-associated factors (Fig. 3). The first approach was to immunoprecipitate Six1 and associated proteins using a Flag-tag appended to either the N- or C-terminus of the Six1 protein. The second approach used proximity ligation to biochemically label proteins associated with Six1 in cellulo (BioID2; Kim et al., 2016); in this case a biotin ligase from A. aeolicus was fused to either the N- or C-terminus of Six1. All of the Six1 constructs were transiently expressed in human HEK293T cells, protein extracted and the presence of Six1 confirmed by Western and the presence of biotinylated proteins in the relevant samples using Neutravidin HRP. Six1 was purified using an antibody to the Flag-tag bound to beads, whereas proteins biotinylated by the Six1-BioID2 were purified using streptavidin magnetic beads. Purified proteins were digested directly on the beads using Trypsin and peptides were separated by LC/MS/MS using a Thermo Orbitrap fusion tribrid. Spectra were analyzed against the human proteome using Scaffold5 proteome software. Using Eya1, a known partner of Six1, and the novel proteins identified by the Y2H screen described above, we analyzed the results using two different strategies. The first strategy was to select only proteins with at least 3 unique high confidence peptides (>95%). The benefit of this method is that each protein identified is clearly present in the sample; the drawback is that abundant proteins are more likely to be detected at the expense of proteins expressed at low levels such as transcription factors. This approach gave us a manageable list of proteins that could interact with Six1, but one that did not overlap with any of the previously identified candidates. Our second strategy was to select proteins identified in at least two independent experiments - BioID (2 replicates) and Six1 immunoprecipitation (18 experiments) with a minimum of one unique high confidence peptide (yielding 684 proteins). After cross-referencing these with genes expressed in Xenopus cranial placodes (Xenbase Xenopus Anatomy Ontology: 2465 proteins), we identified 81 proteins, including 10 candidates (Eya3, Gmeb2, Kdm6b, Mcrs1, Myo18a, Rpl13a, Rps27, Sobp, Tle2, Tmcc3) that were found in the Xenopus Gene Collection and Y2H screens described above.

Figure 3:

Overall view of the Mass spectrometry experiments. Tagged Six1 constructs were transfected into HEK293T cells. Proteins were harvested, the expression level of Six1 was evaluated by western blot (WB) and proteins associated with Six1 immunoprecipitated using either the anti-Flag-tag antibody (left side) or streptavidin beads (right side) for the Six1 Bio-ID2 transfections. All purified proteins were digested with trypsin and peptides separated and identified by LC/Ms/Ms. Protein analysis was performed using two different criteria based on the number of unique high confidence peptides (either 3 peptides or a 1 minimum of 1 peptide). None of the previously identified partners of Six1 were found on the list from the 3 unique peptides, but were found in the list using 1 minimum peptide. The list of putative candidates was compared to genes expressed in Xenopus placodes (81 proteins in Table 4) and the common proteins were further analyzed using a String network analysis (Figure 4).

We used this list of 81 putative Six1 interactors (Table 4) to search the String protein database (https://version-11-5.string-db.org/cgi/network?networkId=b3V8l02dS8mr) and found statistical enrichment in pathways related to hearing and deafness, RNA biogenesis and regulation of transcription as well as development of sensory organs (Fig. 4). We also found a strong link to Sonic hedgehog (Shh) and Wnt signaling pathways. Among these candidates, we prioritized proteins with known roles in the regulation of gene expression (proteins underlined and bold in Table 4), as these are the most likely to modulate Six1 transcriptional activity, such as the kinase Aurkb and the ubiquitin ligase CBL, or that might affect its subcellular localization (e.g., Myosin).

Table 4: Putative Xenopus Six1-interacting proteins identified by LC/MS/MS.

From two different proteomic screens, those identified by one unique high confidence peptide and that are expressed in Xenopus cranial placodes were cross-referenced. Protein Name, String ID and annotation are provided. Proteins with known roles in the regulation of transcription are in bold and underlined.

1	ADAMTS1	9606.ENSP00000284984	A disintegrin and metalloproteinase with thrombospondin motifs 1; Cleaves aggrecan, a cartilage proteoglycan, at the '1938-Glu-∣-Leu-1939' site (within the chondroitin sulfate attachment domain), and may be involved in its turnover (By similarity). Has angiogenic inhibitor activity. Active metalloprotease, which may be associated with various inflammatory processes as well as development of cancer cachexia. May play a critical role in follicular rupture.
2	AHR	9606.ENSP00000242057	Aryl hydrocarbon receptor; Ligand-activated transcriptional activator. Binds to the XRE promoter region of genes it activates. Activates the expression of multiple phase I and II xenobiotic chemical metabolizing enzyme genes (such as the CYP1A1 gene). Mediates biochemical and toxic effects of halogenated aromatic hydrocarbons. Involved in cell-cycle regulation. Likely to play an important role in the development and maturation of many tissues.
3	AR	9606.ENSP00000363822	Androgen receptor; Steroid hormone receptors are ligand-activated transcription factors that regulate eukaryotic gene expression and affect cellular proliferation and differentiation in target tissues. Transcription factor activity is modulated by bound coactivator and corepressor proteins.
4	ATOH1	9606.ENSP00000302216	Protein atonal homolog 1; Transcriptional regulator. Activates E box-dependent transcription in collaboration with TCF3/E47, but the activity is completely antagonized by the negative regulator of neurogenesis HES1. Plays a role in the differentiation of subsets of neural cells by activating E box-dependent transcription; Basic helix-loop-helix proteins.
5	AURKB	9606.ENSP00000313950	Aurora kinase B; Serine/threonine-protein kinase component of the chromosomal passenger complex (CPC), a complex that acts as a key regulator of mitosis. The CPC complex has essential functions at the centromere in ensuring correct chromosome alignment and segregation and is required for chromatin-induced microtubule stabilization and spindle assembly. Involved in the bipolar attachment of spindle microtubules to kinetochores and is a key regulator for the onset of cytokinesis during mitosis. Required for central/midzone spindle assembly and cleavage furrow formation.
6	C5	9606.ENSP00000223642	Complement C5; Activation of C5 by a C5 convertase initiates the spontaneous assembly of the late complement components, C5-C9, into the membrane attack complex. C5b has a transient binding site for C6. The C5b-C6 complex is the foundation upon which the lytic complex is assembled.
7	CACNA1C	9606.ENSP00000266376	Voltage-dependent L-type calcium channel subunit alpha-1C; Voltage-sensitive calcium channels (VSCC) mediate the entry of calcium ions into excitable cells and are also involved in a variety of calcium-dependent processes, including muscle contraction, hormone or neurotransmitter release, gene expression, cell motility, cell division and cell death. The isoform alpha-1C gives rise to L-type calcium currents. Long-lasting (L-type) calcium channels belong to the 'high-voltage activated' (HVA) group.
8	CASR	9606.ENSP00000420194	Extracellular calcium-sensing receptor; G-protein-coupled receptor that senses changes in the extracellular concentration of calcium ions and plays a key role in maintaining calcium homeostasis. Senses fluctuations in the circulating calcium concentration and modulates the production of parathyroid hormone (PTH) in parathyroid glands (By similarity). The activity of this receptor is mediated by a G-protein that activates a phosphatidylinositol- calcium second messenger system. The G-protein-coupled receptor activity is activated by a co-agonist mechanism.
9	CBFA2T2	9606.ENSP00000262653	Protein CBFA2T2; Transcriptional corepressor which facilitates transcriptional repression via its association with DNA-binding transcription factors and recruitment of other corepressors and histone-modifying enzymes. Via association with PRDM14 is involved in regulation of embryonic stem cell (ESC) pluripotency. Involved in primordial germ cell (PCG) formation. Stabilizes PRDM14 and OCT4 on chromatin in a homooligomerization-dependent manner. Can repress the expression of MMP7 in a ZBTB33-dependent manner.
10	CBL	9606.ENSP00000264033	E3 ubiquitin-protein ligase CBL; Adapter protein that functions as a negative regulator of many signaling pathways that are triggered by activation of cell surface receptors. Acts as an E3 ubiquitin-protein ligase, which accepts ubiquitin from specific E2 ubiquitin-conjugating enzymes, and then transfers it to substrates promoting their degradation by the proteasome. Recognizes activated receptor tyrosine kinases, including KIT, FLT1, FGFR1, FGFR2, PDGFRA, PDGFRB, EGFR, CSF1R, EPHA8 and KDR and terminates signaling. Recognizes membrane-bound HCK, SRC and other kinases of the SRC family.
11	CD44	9606.ENSP00000398632	CD44 antigen; Receptor for hyaluronic acid (HA). Mediates cell-cell and cell-matrix interactions through its affinity for HA, and possibly also through its affinity for other ligands such as osteopontin, collagens, and matrix metalloproteinases (MMPs). Adhesion with HA plays an important role in cell migration, tumor growth and progression. In cancer cells, may play an important role in invadopodia formation. Also involved in lymphocyte activation, recirculation and homing, and in hematopoiesis. Altered expression or dysfunction causes numerous pathogenic phenotypes.
12	CHD7	9606.ENSP00000392028	Chromodomain-helicase-DNA-binding protein 7; Probable transcription regulator. Maybe involved in the in 45S precursor rRNA production; DNA helicases.
13	CHMP4B	9606.ENSP00000217402	Charged multivesicular body protein 4b; Probable core component of the endosomal sorting required for transport complex III (ESCRT-III) which is involved in multivesicular bodies (MVBs) formation and sorting of endosomal cargo proteins into MVBs. MVBs contain intraluminal vesicles (ILVs) that are generated by invagination and scission from the limiting membrane of the endosome and mostly are delivered to lysosomes enabling degradation of membrane proteins, such as stimulated growth factor receptors, lysosomal enzymes and lipids.
14	CLCNKB	9606.ENSP00000364831	Chloride channel protein ClC-Kb; Voltage-gated chloride channel. Chloride channels have several functions including the regulation of cell volume; membrane potential stabilization, signal transduction and transepithelial transport. May be important in urinary concentrating mechanisms; Belongs to the chloride channel (TC 2.A.49) family. CLCNKB subfamily
15	CNN1	9606.ENSP00000252456	Calponin-1; Thin filament-associated protein that is implicated in the regulation and modulation of smooth muscle contraction. It is capable of binding to actin, calmodulin, troponin C and tropomyosin. The interaction of calponin with actin inhibits the actomyosin Mg-ATPase activity (By similarity)
16	CSNK1G2	9606.ENSP00000255641	Casein kinase I isoform gamma-2; Serine/threonine-protein kinase. Casein kinases are operationally defined by their preferential utilization of acidic proteins such as caseins as substrates. It can phosphorylate a large number of proteins. Participates in Wnt signaling. Phosphorylates COL4A3BP/CERT, MTA1 and SMAD3. Involved in brain development and vesicular trafficking and neurotransmitter releasing from small synaptic vesicles. Regulates fast synaptic transmission mediated by glutamate.
17	DDX21	9606.ENSP00000346120	Nucleolar RNA helicase 2; RNA helicase that acts as a sensor of the transcriptional status of both RNA polymerase (Pol) I and II: promotes ribosomal RNA (rRNA) processing and transcription from polymerase II (Pol II). Binds various RNAs, such as rRNAs, snoRNAs, 7SK and, at lower extent, mRNAs. In the nucleolus, localizes to rDNA locus, where it directly binds rRNAs and snoRNAs, and promotes rRNA transcription, processing and modification. Required for rRNA 2'- O-methylation, possibly by promoting the recruitment of late-acting snoRNAs SNORD56 and SNORD58 with pre-ribosomal complexes.
18	DDX43	9606.ENSP00000359361	Probable ATP-dependent RNA helicase DDX43; DEAD-box helicase 43
19	DDX59	9606.ENSP00000330460	Probable ATP-dependent RNA helicase DDX59; Zinc fingers HIT-type; Belongs to the DEAD box helicase family. DDX59 subfamily
20	DHRS7	9606.ENSP00000216500	Dehydrogenase/reductase SDR family member 7; Short chain dehydrogenase/reductase superfamily
21	DICER1	9606.ENSP00000437256	Dicer 1, ribonuclease iii; Endoribonuclease Dicer; Double-stranded RNA (dsRNA) endoribonuclease playing a central role in short dsRNA-mediated post-transcriptional gene silencing. Cleaves naturally occurring long dsRNAs and short hairpin pre-microRNAs (miRNA) into fragments of twenty-one to twenty-three nucleotides with 3' overhang of two nucleotides, producing respectively short interfering RNAs (siRNA) and mature microRNAs. SiRNAs and miRNAs serve as guide to direct the RNA-induced silencing complex (RISC) to complementary RNAs to degrade them or prevent their translation.
22	DTL	9606.ENSP00000355958	Denticleless protein homolog; Substrate-specific adapter of a DCX (DDB1-CUL4-X-box) E3 ubiquitin-protein ligase complex required for cell cycle control, DNA damage response and translesion DNA synthesis. The DCX(DTL) complex, also named CRL4(CDT2) complex, mediates the polyubiquitination and subsequent degradation of CDT1, CDKN1A/p21(CIP1), FBXO18/FBH1, KMT5A and SDE2. CDT1 degradation in response to DNA damage is necessary to ensure proper cell cycle regulation of DNA replication.
23	EFNB1	9606.ENSP00000204961	Ephrin-B1; Binds to the receptor tyrosine kinases EPHB1 and EPHA1. Binds to, and induce the collapse of, commissural axons/growth cones in vitro. May play a role in constraining the orientation of longitudinally projecting axons.
24	ELAVL1	9606.ENSP00000385269	ELAV-like protein 1; RNA-binding protein that binds to the 3'-UTR region of mRNAs and increases their stability. Involved in embryonic stem cells (ESCs) differentiation: preferentially binds mRNAs that are not methylated by N6-methyladenosine (m6A), stabilizing them, promoting ESCs differentiation (By similarity). Binds to poly-U elements and AU-rich elements (AREs) in the 3'-UTR of target mRNAs. Binds avidly to the AU-rich element in FOS and IL3/interleukin-3 mRNAs.
25	EPN3	9606.ENSP00000268933	Epsin-3; Epsin 3; Belongs to the epsin family
26	EPPK1	9606.ENSP00000484472	Epiplakin; Cytoskeletal linker protein that connects to intermediate filaments and controls their reorganization in response to stress. In response to mechanical stress like wound healing, is associated with the machinery for cellular motility by slowing down keratinocyte migration and proliferation and accelerating keratin bundling in proliferating keratinocytes thus contributing to tissue architecture.
27	EPS8	9606.ENSP00000281172	Epidermal growth factor receptor kinase substrate 8; Signaling adapter that controls various cellular protrusions by regulating actin cytoskeleton dynamics and architecture. Depending on its association with other signal transducers, can regulate different processes. Together with SOS1 and ABI1, forms a trimeric complex that participates in transduction of signals from Ras to Rac by activating the Rac- specific guanine nucleotide exchange factor (GEF) activity. Acts as a direct regulator of actin dynamics by binding actin filaments.
28	ESR1	9606.ENSP00000405330	Estrogen receptor; Nuclear hormone receptor. The steroid hormones and their receptors are involved in the regulation of eukaryotic gene expression and affect cellular proliferation and differentiation in target tissues. Ligand-dependent nuclear transactivation involves either direct homodimer binding to a palindromic estrogen response element (ERE) sequence or association with other DNA- binding transcription factors, such as AP-1/c-Jun, c-Fos, ATF-2, Sp1 and Sp3, to mediate ERE-independent signaling.
29	EYA3	9606.ENSP00000362978	Eyes absent homolog 3; Tyrosine phosphatase that specifically dephosphorylates 'Tyr-142' of histone H2AX (H2AXY142ph). 'Tyr-142' phosphorylation of histone H2AX plays a central role in DNA repair and acts as a mark that distinguishes between apoptotic and repair responses to genotoxic stress. Promotes efficient DNA repair by dephosphorylating H2AX, promoting the recruitment of DNA repair complexes containing MDC1. Its function as histone phosphatase probably explains its role in transcription regulation during organogenesis. Coactivates SIX1, and seems to coactivate SIX2, SIX4 and SIX5.
30	FAS	9606.ENSP00000347979	Tumor necrosis factor receptor superfamily member 6; Receptor for TNFSF6/FASLG. The adapter molecule FADD recruits caspase-8 to the activated receptor. The resulting death- inducing signaling complex (DISC) performs caspase-8 proteolytic activation which initiates the subsequent cascade of caspases (aspartate-specific cysteine proteases) mediating apoptosis. FAS-mediated apoptosis may have a role in the induction of peripheral tolerance, in the antigen-stimulated suicide of mature T-cells, or both. The secreted isoforms 2 to 6 block apoptosis in vitro.
31	FLOT2	9606.ENSP00000378368	Flotillin-2; May act as a scaffolding protein within caveolar membranes, functionally participating in formation of caveolae or caveolae-like vesicles. May be involved in epidermal cell adhesion and epidermal structure and function; Belongs to the band 7/mec-2 family.
32	FLVCR2	9606.ENSP00000238667	Feline leukemia virus subgroup C receptor-related protein 2; Acts as an importer of heme. Also acts as a transporter for a calcium-chelator complex, important for growth and calcium metabolism.
33	GLI2	9606.ENSP00000390436	Zinc finger protein GLI2; Functions as transcription regulator in the hedgehog (Hh) pathway. Functions as transcriptional activator. May also function as transcriptional repressor (By similarity). Requires STK36 for full transcriptional activator activity. Required for normal embryonic development; Zinc fingers C2H2-type
34	GLI3	9606.ENSP00000379258	Transcriptional activator GLI3; Has a dual function as a transcriptional activator and a repressor of the sonic hedgehog (Shh) pathway, and plays a role in limb development. The full-length GLI3 form (GLI3FL) after phosphorylation and nuclear translocation, acts as an activator (GLI3A) while GLI3R, its C-terminally truncated form, acts as a repressor. A proper balance between the GLI3 activator and the repressor GLI3R, rather than the repressor gradient itself or the activator/repressor ratio gradient, specifies limb digit number and identity.
35	GNL3	9606.ENSP00000395772	Guanine nucleotide-binding protein-like 3; May be required to maintain the proliferative capacity of stem cells. Stabilizes MDM2 by preventing its ubiquitination, and hence proteasomal degradation (By similarity); Belongs to the TRAFAC class YlqF/YawG GTPase family.
36	ID4	9606.ENSP00000367972	DNA-binding protein inhibitor ID-4; Transcriptional regulator (lacking a basic DNA binding domain) which negatively regulates the basic helix-loop-helix (bHLH) transcription factors by forming heterodimers and inhibiting their DNA binding and transcriptional activity. Implicated in regulating a variety of cellular processes, including cellular growth, senescence, differentiation, apoptosis, angiogenesis, and neoplastic transformation.
37	IL17RA	9606.ENSP00000320936	Interleukin-17 receptor A; Receptor for IL17A. Receptor for IL17F. Binds to IL17A with higher affinity than to IL17F. Binds IL17A and IL17F homodimers as part of a heterodimeric complex with IL17RC. Also binds heterodimers formed by IL17A and IL17F as part of a heterodimeric complex with IL17RC. Receptor for IL17C as part of a heterodimeric complex with IL17RE. Activation of IL17RA leads to induction of expression of inflammatory chemokines and cytokines such as CXCL1, CXCL8/IL8 and IL6.
38	IL6ST	9606.ENSP00000370698	Interleukin-6 receptor subunit beta; Signal-transducing molecule. The receptor systems for IL6, LIF, OSM, CNTF, IL11, CTF1 and BSF3 can utilize IL6ST for initiating signal transmission. Binding of IL6 to IL6R induces IL6ST homodimerization and formation of a high-affinity receptor complex, which activates Janus kinases. That causes phosphorylation of IL6ST tyrosine residues which in turn activates STAT3. Mediates signals which regulate immune response, hematopoiesis, pain control and bone metabolism and has a role in embryonic development.
39	IRS1	9606.ENSP00000304895	Insulin receptor substrate 1; May mediate the control of various cellular processes by insulin. When phosphorylated by the insulin receptor binds specifically to various cellular proteins containing SH2 domains such as phosphatidylinositol 3-kinase p85 subunit or GRB2. Activates phosphatidylinositol 3-kinase when bound to the regulatory p85 subunit.
40	KCNQ1	9606.ENSP00000155840	Potassium voltage-gated channel subfamily KQT member 1; Potassium channel that plays an important role in a number of tissues, including heart, inner ear, stomach and colon (By similarity). Associates with KCNE beta subunits that modulates current kinetics (By similarity). Induces a voltage-dependent by rapidly activating and slowly deactivating potassium-selective outward current (By similarity). Promotes also a delayed voltage activated potassium current showing outward rectification characteristic (By similarity).
41	KDM6B	9606.ENSP00000254846	Lysine-specific demethylase 6B; Histone demethylase that specifically demethylates 'Lys-27' of histone H3, thereby playing a central role in histone code. Demethylates trimethylated and dimethylated H3 'Lys-27'. Plays a central role in regulation of posterior development, by regulating HOX gene expression. Involved in inflammatory response by participating in macrophage differentiation in case of inflammation by regulating gene expression and macrophage differentiation. Plays a demethylase-independent role in chromatin remodeling to regulate T-box family member-dependent gene expression.
42	KIF2C	9606.ENSP00000361298	Kinesin-like protein KIF2C; In complex with KIF18B, constitutes the major microtubule plus-end depolymerizing activity in mitotic cells. Regulates the turnover of microtubules at the kinetochore and functions in chromosome segregation during mitosis. Plays a role in chromosome congression and is required for the lateral to end-on conversion of the chromosome- microtubule attachment; Belongs to the TRAFAC class myosin-kinesin ATPase superfamily. Kinesin family. MCAK/KIF2 subfamily.
43	LRSAM1	9606.ENSP00000322937	Leucine rich repeat and sterile alpha motif containing 1; E3 ubiquitin-protein ligase LRSAM1; E3 ubiquitin-protein ligase that mediates monoubiquitination of TSG101 at multiple sites, leading to inactivate the ability of TSG101 to sort endocytic (EGF receptors) and exocytic (HIV-1 viral proteins) cargos. Bacterial recognition protein that defends the cytoplasm from invasive pathogens. Localizes to several intracellular bacterial pathogens and generates the bacteria-associated ubiquitin signal leading to autophagy-mediated intracellular bacteria degradation (xenophagy); Ring finger proteins.
44	LRTOMT	9606.ENSP00000305742	Transmembrane O-methyltransferase; Catalyzes the O-methylation, and thereby the inactivation, of catecholamine neurotransmitters and catechol hormones (By similarity). Required for auditory function. Component of the cochlear hair cell's mechanotransduction (MET) machinery. Involved in the assembly of the asymmetric tip-link MET complex. Required for transportation of TMC1 and TMC2 proteins into the mechanically sensitive stereocilia of the hair cells.
45	MAPK15	9606.ENSP00000337691	Mitogen-activated protein kinase 15; In vitro, phosphorylates MBP; Mitogen-activated protein kinases
46	MTNR1A	9606.ENSP00000302811	Melatonin receptor type 1A; High affinity receptor for melatonin. Likely to mediate the reproductive and circadian actions of melatonin. The activity of this receptor is mediated by pertussis toxin sensitive G proteins that inhibit adenylate cyclase activity; Belongs to the G-protein coupled receptor 1 family.
47	MYO3A	9606.ENSP00000265944	Myosin-IIIa; Probable actin-based motor with a protein kinase activity. Probably plays a role in vision and hearing. Required for normal cochlear hair bundle development and hearing. Plays an important role in the early steps of cochlear hair bundle morphogenesis. Influences the number and lengths of stereocilia to be produced and limits the growth of microvilli within the forming auditory hair bundles thereby contributing to the architecture of the hair bundle, including its staircase pattern.
48	MYO6	9606.ENSP00000358994	Unconventional myosin-VI; Myosins are actin-based motor molecules with ATPase activity. Unconventional myosins serve in intracellular movements. Myosin 6 is a reverse-direction motor protein that moves towards the minus-end of actin filaments. Has slow rate of actin-activated ADP release due to weak ATP binding. Functions in a variety of intracellular processes such as vesicular membrane trafficking and cell migration. Required for the structural integrity of the Golgi apparatus via the p53-dependent pro-survival pathway.
49	NEFM	9606.ENSP00000221166	Neurofilament medium polypeptide (neurofilament 3); Neurofilament medium polypeptide; Neurofilaments usually contain three intermediate filament proteins: L, M, and H which are involved in the maintenance of neuronal caliber.
50	PCNA	9606.ENSP00000368458	Proliferating cell nuclear antigen; Auxiliary protein of DNA polymerase delta and is involved in the control of eukaryotic DNA replication by increasing the polymerase's processibility during elongation of the leading strand. Induces a robust stimulatory effect on the 3'- 5' exonuclease and 3'-phosphodiesterase, but not apurinic- apyrimidinic (AP) endonuclease, APEX2 activities. Has to be loaded onto DNA in order to be able to stimulate APEX2. Plays a key role in DNA damage response (DDR) by being conveniently positioned at the replication fork to coordinate DNA replication.
51	PFN2	9606.ENSP00000239940	Profilin-2; Binds to actin and affects the structure of the cytoskeleton. At high concentrations, profilin prevents the polymerization of actin, whereas it enhances it at low concentrations. By binding to PIP2, it inhibits the formation of IP3 and DG.
52	PRDM16	9606.ENSP00000270722	PR domain zinc finger protein 16; Binds DNA and functions as a transcriptional regulator. Functions in the differentiation of brown adipose tissue (BAT) which is specialized in dissipating chemical energy in the form of heat in response to cold or excess feeding while white adipose tissue (WAT) is specialized in the storage of excess energy and the control of systemic metabolism. Together with CEBPB, regulates the differentiation of myoblastic precursors into brown adipose cells. Functions also as a repressor of TGF-beta signaling.
53	PRKACA	9606.ENSP00000309591	cAMP-dependent protein kinase catalytic subunit alpha; Phosphorylates a large number of substrates in the cytoplasm and the nucleus. Regulates the abundance of compartmentalized pools of its regulatory subunits through phosphorylation of PJA2 which binds and ubiquitinates these subunits, leading to their subsequent proteolysis. Phosphorylates CDC25B, ABL1, NFKB1, CLDN3, PSMC5/RPT6, PJA2, RYR2, RORA and VASP. RORA is activated by phosphorylation. Required for glucose- mediated adipogenic differentiation increase and osteogenic differentiation inhibition from osteoblasts.
54	Protein	StringId	Annotation
55	PSMD6	9606.ENSP00000418695	Proteasome 26S subunit, non-ATPase 6
56	RAN	9606.ENSP00000446215	GTP-binding nuclear protein Ran; GTPase involved in nucleocytoplasmic transport, participating both to the import and the export from the nucleus of proteins and RNAs. Switches between a cytoplasmic GDP- and a nuclear GTP-bound state by nucleotide exchange and GTP hydrolysis. Nuclear import receptors such as importin beta bind their substrates only in the absence of GTP-bound RAN and release them upon direct interaction with GTP-bound RAN while export receptors behave in the opposite way.
57	REST	9606.ENSP00000311816	RE1-silencing transcription factor; Transcriptional repressor which binds neuron-restrictive silencer element (NRSE) and represses neuronal gene transcription in non-neuronal cells. Restricts the expression of neuronal genes by associating with two distinct corepressors, mSin3 and CoREST, which in turn recruit histone deacetylase to the promoters of REST-regulated genes. Mediates repression by recruiting the BHC complex at RE1/NRSE sites which acts by deacetylating and demethylating specific sites on histones, thereby acting as a chromatin modifier..
58	RPL19	9606.ENSP00000225430	Large subunit ribosomal protein l19e; Ribosomal protein L19; Belongs to the eukaryotic ribosomal protein eL19 family.
59	RPL23	9606.ENSP00000420311	Large subunit ribosomal protein l23e; Ribosomal protein L23.
60	RPL29	9606.ENSP00000418868	60S ribosomal protein L29; Component of the large ribosomal subunit.
61	RPS11	9606.ENSP00000270625	Small subunit ribosomal protein s11e; Ribosomal protein S11; Belongs to the universal ribosomal protein uS17 family.
62	RPS3A	9606.ENSP00000346050	40S ribosomal protein S3a; May play a role during erythropoiesis through regulation of transcription factor DDIT3; Belongs to the eukaryotic ribosomal protein eS1 family.
63	RUNX2	9606.ENSP00000360493	Runt-related transcription factor 2; Transcription factor involved in osteoblastic differentiation and skeletal morphogenesis. Essential for the maturation of osteoblasts and both intramembranous and endochondral ossification. CBF binds to the core site, 5'-PYGPYGGT-3', of a number of enhancers and promoters, including murine leukemia virus, polyomavirus enhancer, T-cell receptor enhancers, osteocalcin, osteopontin, bone sialoprotein, alpha 1(I) collagen, LCK, IL-3 and GM-CSF promoters. In osteoblasts, supports transcription activation.
64	SAG	9606.ENSP00000386444	S-arrestin; Arrestin is one of the major proteins of the ros (retinal rod outer segments); it binds to photoactivated- phosphorylated rhodopsin, thereby apparently preventing the transducin-mediated activation of phosphodiesterase.
65	SALL4	9606.ENSP00000217086	Sal-like protein 4; Transcription factor with a key role in the maintenance and self-renewal of embryonic and hematopoietic stem cells; Zinc fingers C2H2-type.
66	SCRIB	9606.ENSP00000349486	Protein scribble homolog; Scaffold protein involved in different aspects of polarized cells differentiation regulating epithelial and neuronal morphogenesis. Most probably functions in the establishment of apico-basal cell polarity. May function in cell proliferation regulating progression from G1 to S phase and as a positive regulator of apoptosis for instance during acinar morphogenesis of the mammary epithelium. May also function in cell migration and adhesion and hence regulate cell invasion through MAPK signaling. May play a role in exocytosis and in the targeting synaptic vesicles.
67	SHH	9606.ENSP00000297261	Sonic hedgehog protein; Sonic hedgehog protein: The C-terminal part of the sonic hedgehog protein precursor displays an autoproteolysis and a cholesterol transferase activity (By similarity). Both activities result in the cleavage of the full-length protein into two parts (ShhN and ShhC) followed by the covalent attachment of a cholesterol moiety to the C-terminal of the newly generated ShhN (By similarity). Both activities occur in the reticulum endoplasmic (By similarity). Once cleaved, ShhC is degraded in the endoplasmic reticulum (By similarity); Hedgehog signaling molecule family.
68	SHROOM2	9606.ENSP00000370299	Protein Shroom2; May be involved in endothelial cell morphology changes during cell spreading. In the retinal pigment epithelium, may regulate the biogenesis of melanosomes and promote their association with the apical cell surface by inducing gamma-tubulin redistribution (By similarity); Belongs to the shroom family.
69	SHROOM4	9606.ENSP00000365188	Protein Shroom4; Probable regulator of cytoskeletal architecture that plays an important role in development. May regulate cellular and cytoskeletal architecture by modulating the spatial distribution of myosin II (By similarity); Belongs to the shroom family.
70	Six1	9606.ENSP00000247182	Homeobox protein SIX1; Transcription factor that is involved in the regulation of cell proliferation, apoptosis and embryonic development. Plays an important role in the development of several organs, including kidney, muscle and inner ear. Depending on context, functions as transcriptional repressor or activator. Lacks an activation domain, and requires interaction with EYA family members for transcription activation. Mediates nuclear translocation of EYA1 and EYA2. Binds the 5'-TCA[AG][AG]TTNC-3' motif present in the MEF3 element in the MYOG promoter.
71	SIX4	9606.ENSP00000216513	Homeobox protein SIX4; Transcriptional regulator which can act as both a transcriptional repressor and activator by binding a DNA sequence on these target genes and is involved in processes like cell differentiation, cell migration and cell survival. Transactivates gene expression by binding a 5'-[CAT]A[CT][CT][CTG]GA[GAT]-3' motif present in the Trex site and a 5'-TCA[AG][AG]TTNC-3' motif present in the MEF3 site of the muscle-specific genes enhancer. Acts cooperatively with EYA proteins to transactivate their target genes through interaction and nuclear translocation of EYA protein.
72	SKA3	9606.ENSP00000319417	Spindle and kinetochore-associated protein 3; Component of the SKA1 complex, a microtubule-binding subcomplex of the outer kinetochore that is essential for proper chromosome segregation. The SKA1 complex is a direct component of the kinetochore-microtubule interface and directly associates with microtubules as oligomeric assemblies. The complex facilitates the processive movement of microspheres along a microtubule in a depolymerization-coupled manner. In the complex, it mediates the microtubule-stimulated oligomerization.
73	SLC12A2	9606.ENSP00000262461	Solute carrier family 12 member 2; Electrically silent transporter system. Mediates sodium and chloride reabsorption. Plays a vital role in the regulation of ionic balance and cell volume; Belongs to the SLC12A transporter family.
74	SLC22A7	9606.ENSP00000361666	Solute carrier family 22 member 7; Mediates sodium-independent multispecific organic anion transport. Transport of prostaglandin E2, prostaglandin F2, tetracycline, bumetanide, estrone sulfate, glutarate, dehydroepiandrosterone sulfate, allopurinol, 5-fluorouracil, paclitaxel, L-ascorbic acid, salicylate, ethotrexate, and alpha- ketoglutarate; Solute carriers.
75	SLC2A1	9606.ENSP00000416293	Solute carrier family 2, facilitated glucose transporter member 1; Facilitative glucose transporter. This isoform may be responsible for constitutive or basal glucose uptake. Has a very broad substrate specificity; can transport a wide range of aldoses including both pentoses and hexoses; Belongs to the major facilitator superfamily. Sugar transporter (TC 2.A.1.1) family. Glucose transporter subfamily.
76	SLIT1	9606.ENSP00000266058	Slit guidance ligand 1; Slit homolog 1 protein; Thought to act as molecular guidance cue in cellular migration, and function appears to be mediated by interaction with roundabout homolog receptors. During neural development involved in axonal navigation at the ventral midline of the neural tube and projection of axons to different regions (By similarity). SLIT1 and SLIT2 together seem to be essential for midline guidance in the forebrain by acting as repulsive signal preventing inappropriate midline crossing by axons projecting from the olfactory bulb
77	TF	9606.ENSP00000385834	Serotransferrin; Transferrins are iron binding transport proteins which can bind two Fe(3+) ions in association with the binding of an anion, usually bicarbonate. It is responsible for the transport of iron from sites of absorption and heme degradation to those of storage and utilization. Serum transferrin may also have a further role in stimulating cell proliferation.
78	TLE1	9606.ENSP00000365682	Transducin-like enhancer protein 1; Transcriptional corepressor that binds to a number of transcription factors. Inhibits NF-kappa-B-regulated gene expression. Inhibits the transcriptional activation mediated by FOXA2, and by CTNNB1 and TCF family members in Wnt signaling. The effects of full-length TLE family members may be modulated by association with dominant-negative AES. Unusual function as coactivator for ESRRG; Belongs to the WD repeat Groucho/TLE family.
79	TRIM29	9606.ENSP00000343129	Tripartite motif-containing protein 29; It is able to complement the radiosensitivity defect of an ataxia telangiectasia (AT) fibroblast cell line; Tripartite motif containing.
80	TRPC3	9606.ENSP00000368966	Short transient receptor potential channel 3; Thought to form a receptor-activated non-selective calcium permeant cation channel. Probably is operated by a phosphatidylinositol second messenger system activated by receptor tyrosine kinases or G-protein coupled receptors. Activated by diacylglycerol (DAG) in a membrane-delimited fashion, independently of protein kinase C, and by inositol 1,4,5- triphosphate receptors (ITPR) with bound IP3. May also be activated by internal calcium store depletion; Transient receptor potential cation channels.
81	TRPM3	9606.ENSP00000366314	Transient receptor potential cation channel subfamily M member 3; Calcium channel mediating constitutive calcium ion entry. Its activity is increased by reduction in extracellular osmolarity, by store depletion and muscarinic receptor activation; Belongs to the transient receptor (TC 1.A.4) family. LTrpC subfamily. TRPM3 sub-subfamily.
82	ZMYM3	9606.ENSP00000322845	Zinc finger MYM-type protein 3; Plays a role in the regulation of cell morphology and cytoskeletal organization; Zinc fingers MYM-type.

Figure 4: Proteins identified by LC/MS/MS and present in Xenopus cranial placodes.

A. Putative interactions were analyzed using the String web site (https://version-11-5.string-db.org/cgi/network?networkId=b3V8l02dS8mr). Proteins with known physical or functional interactions are linked.

B. Proteins with roles in otic development, hearing loss, nuclear localization or transcriptional regulation are color coded.

What is most interesting from our combined proteomic analysis is the overlap we discovered with previous screens described above. For instance, the transcriptomic studies also identified Atoh1 and Shh and the Xenopus Gene Collection and Y2H screens also identified 10 of our top candidates. We have thus identified a large number of candidate proteins that need to be investigated for their developmental functions in otic development and potential roles in hearing loss. Further analysis of the results from these diverse yet similarly focused screens will help prioritize these candidates for follow-up studies by the community.

VI. Multiplicity of Six Complexes and Relevance to Disease Etiology

Given the direct association of both SIX1 and EYA1 with BOS/BOR, much of the research on pathological variants has focused on this complex. However, most SIX1 variants do not appear to be defective in their interaction with EYA (Table 1) and the few that do have not been queried vis-a-vis other co-factors. Moreover, interactions of SIX1 with some of the other co-factors involved in otic development occur in competition with EYA1 and cofactor-cofactor interactions also influence complex function. As shown by Tavares and colleagues (2021), Sobp interacts with both Six1 and Eya1, yet it has differential effects on their respective abilities to enter the nucleus. Thus, it is imperative that our thinking moves beyond simple binary complexes (Six1 + co-factor) and consider additional levels of complexity that are likely at play under physiological conditions, and in the etiology of BOS/BOR.

To begin such an analysis, the Pignoni laboratory used a Y2H approach to examine the ability of Xenopus Six1 (SD+HD) fragments harboring known BOS/BOR variants to interact with three established co-factors (Eya1, Tle4 and D-Sobp; see below regarding use of the fly Sobp fragment). The results were intriguing; in this assay, many of the clinical variants did not affect co-factor binding, and among those that did, none of the mutations specifically affected binding to Eya1 (Fig. 5). The V17E mutation, previously shown to impair SIX1-EYA1 binding (Table 1; Patrick et al., 2009), also abolished interactions with the other two co-factors. The R64H mutation profoundly affected the interactions with Tle4 but only minimally weakened the interaction with D-Sobp and showed no effect on Eya1, and the H73P mutation specifically affected the ability of Six1 to interact with the D-Sobp fragment, with minimal effects on Eya1 and none on Tle4. Lastly, a ‘designer’ mutant, C16R, predicted to disrupt the SIX1-Eya2 binding interface (Patrick et al., 2013) resulted in decreased interaction of Six1 with both Eya1 and Tle4, but not D-Sobp; a different substitution at the same residue, C16S, strongly affected the interaction with Tle4 but was completely or partially tolerated by D-Sobp and Eya1, respectively. These results raise the distinct possibility that some of the syndromic manifestations of BOS/BOR reflect the selective failure of specific SIX1-co-factor interactions, and not necessarily the one with EYA1. Further studies in this direction will need to consider the roles of additional co-factors more critically when experimental results are interpreted.

Figure 5:

Y2H analysis of Six1 (SD+HD) bait fragments (Six1.L, residues 1-194), harboring clinical and designer substitutions within the SD, with prey fragments of Eya1 (Eya1.L, residues 316-587), Tle4 (Tle4.S, residues 1-138) and Drosophila (D)-Sobp (Sobp-PA, residues 210-354). Interactions were assessed via mating of transformed haploid yeast strains (Y2H Gold “bait” + Y187 “prey”) and assessing diploid cell growth across a gradient of 3-AT concentrations (0.1-10 mM) on media lacking His, Ade, Leu and Trp. Semiquantitative interaction scores are color coded as wild type level (magenta), affected interaction (from light to medium green) or no detectable interaction (dark green). They were inferred from the maximum concentration of 3-AT that supported growth, relative to wild type controls, for each interaction pair. The approximate positions of the 6 SD helices are indicated in grey.

To further investigate this possibility, we sought to compare the nature of co-factor-Six1 binding in the Y2H assay. Hence, we performed Alanine-scanning mutagenesis (X>A and A>F) of the Six1 SD and tested the ability of each single-aa mutant SD+HD fragment to interact with fragments of the above three co-factors. This resulted in a positional map of the SD where each residue was assigned a binding score for each co-factor (Fig. 6). As was seen for certain clinical variants, some mutations (e.g., Q13A) primarily affected binding to only one co-factor, Eya1 in this case, whereas mutations in other positions (e.g., G31A) affected binding of all co-factors. In this approach, positive associations with some of the cofactors strongly suggest that the modified protein is stable but defective in the specific interaction, whereas variants affecting binding to all three co-factors may, in some cases, reflect issues with protein stability/expression. More illuminating to us was the fact that the mutations affecting Eya1 binding were clustered in the region encoding the second α-helix of the SD, consistent with the interaction interface revealed by the Six1-Eya2 crystal structure (Patrick et al., 2013), but those affecting interactions with Tle4 and with D-Sobp were broadly distributed across most of the SD. Thus, although binding with TLE factors has also been mapped to the second helix (Zhu, 2002), Six1 binds to this family of co-factors in a way that is distinct from the mechanism at work with Eya1. Furthermore, the fact that Tle4 and Sobp binding are affected by aa substitution across a broad region of the SD, encompassing α-helices 1-5, suggests that these interactions are sensitive to a specific overall conformation of the SD.

Figure 6: Alanine scanning mutagenesis reveals potential cofactor binding sites.

A. Alanine-scanning mutagenesis (X>A, A>F) of the Six1 SD was performed and SD+HD fragments were analyzed for their ability to interact with co-factors in Y2H assays; methods and scoring are as in Figure 5. The approximate positions of the 6 SD helices are indicated in grey.

B. Structural model of the Six1 SD (magenta) and HD (yellow) domains derived from a co-crystal structure with Eya2 (PDB: 4egc)

C-E. Y2H results from (A) were superimposed on the structure (B) to highlight regions (green) where aa substitutions disrupt co-factor interaction. C. Eya1; D. Tle4; E. D-Sobp. C’. Docking of Eya2 along the green surface of Six1. 3D renderings were generated using Chimera (v1.15; UCSF Resource for Biocomputing, Visualization, and Informatics).

As noted above, the SIX1 SD-HD Y2H assays for Sobp were performed with fragments of the Drosophila Sobp protein. This is because we were initially unable to detect an interaction between Six1 and vertebrate (frog or human) Sobp/SOBP in this assay. Yet, the interaction between Six1 and full-length Sobp can be clearly detected by Co-IP from HEK cells (Tavares et al., 2021). In trying to understand this discrepancy, we came upon two intriguing findings. First, since we were primarily interested in protein interactions via the SD, we tested an SD-only fragment of SIX1 in our Y2H assays. This fragment was able to interact with human SOBP. Moreover, this was true for MCRS1 as well, where binding in Y2H assays was detected with the SIX1 SD fragment but not the SD+HD fragment. Second, we discovered that different splice forms of human SOBP encode proteins with differential abilities to interact with SIX factors. Specifically, we isolated an SOBP transcript from a HEK 293T cDNA library that contains an additional exon; this transcript is not presently curated as a known Sobp variant in UCSC Genome Browser (hg38). Interestingly, protein fragments containing this sequence displayed increased binding to the SIX1 SD bait in Y2H assays. We conclude from these observations that the accessibility of certain co-factor binding sites within the SD may be masked and/or regulated by the broader tertiary structure of SIX1. In particular, the impaired binding of SOBP and MCRS1 in the presence of the HD may reflect the lack of a necessary post-translational modification on either factor, a requirement for other portions of the proteins or for a specific DNA-bound state of the HD, and/or the requirement for additional co-factors to promote accessibility of the SD binding surface.

Thus, it is imperative that we begin to consider the roles that higher-order complex structures may play in determining complex composition, and ultimately, in the specificity and processivity of SIX1 transcription. Simple binary interaction assays and/or screens using methods such as Y2H and MS are tremendously powerful for making basic discoveries and for hypothesis generation. However, in vivo analyses in Xenopus will play an essential role in disentangling the nuances of these interactions, and their effects on otic development.

VII. Summary and future directions:

BOS/BOR patients present with variable defects in the inner and middle ears, which are derived from the otic placode and neural crest, respectively; a subset also presents with kidney defects. While genetic screening is considered the best approach to identify newborns at risk for hearing loss and kidney dysfunction, to date only three causative genes - SIX1, SIX5, EYA1 - have been identified and they only account for about half of the patients. Therefore, identifying additional genes that functionally involve Six1, either as a transcriptional target or as an interacting protein in a higher order transcriptional complex, during otic development is likely to uncover new causative genes that can be added to genetic screenings for at-risk newborns. We used the powerful Drosophila model to identify putative BOS/BOR candidate genes, and are now harnessing the biochemical and embryological advantages of Xenopus to determine whether these candidates are functionally required during otic development. Functional testing of these candidates in an aquatic animal model with high genetic, protein and functional similarity to human is likely to rapidly uncover high priority candidates for patient hearing loss genetic screening. With this approach we hope to uncover additional genes that are diagnostic for BOS/BOR and other hearing loss syndromes.

Research Highlights:

BOR patients have hearing loss and kidney defects.

Mutations in SIX1, SIX5 and EYA1 are causative in about half of cases.

To uncover putative new candidate genes, we used Xenopus and identified novel transcriptional targets and interacting proteins.

Data Availability:

All data for this study are presented in the body of the manuscript. Please contact authors for additional details.