Article

ISSN 0006-2979, Biochemistry (Moscow), 2024, Vol. 89, No. 6, pp. 987-1001 © The Author(s) 2024. This article is an open access publication.

987

REVIEW

Irreducible Complexity of Hox Gene:

Path to the Canonical Function of the Hox Cluster

Milana A. Kulakova

1,a

*, Georgy P. Maslakov

, and Liudmila O. Poliushkevich

Department of Embryology, Faculty of Biology, St. Petersburg State University, 199034 St. Petersburg, Russia

e-mail: m.kulakowa@spbu.ru

Received November 19, 2023

Revised March 22, 2024

Accepted March 27, 2024

Abstract— The evolution of major taxa is often associated with the emergence of new gene families. In all mul-

ticellular animals except sponges and comb jellies, the genomes contain Hox genes, which are crucial regula-

tors of development. The canonical function of Hox genes involves colinear patterning of body parts in bilateral

animals. This general function is implemented through complex, precisely coordinated mechanisms, not all of

which are evolutionarily conserved and fully understood. We suggest that the emergence of this regulatory com-

plexity was preceded by a stage of cooperation between more ancient morphogenetic programs or their individual

elements. Footprints of these programs may be present in modern animals to execute non-canonical Hox func-

tions. Non-canonical functions of Hox genes are involved in maintaining terminal nerve cell specificity, autophagy,

oogenesis, pre-gastrulation embryogenesis, vertical signaling, and a number of general biological processes. These

functions are realized by the basic properties of homeodomain protein and could have triggered the evolution of

ParaHoxozoa and Nephrozoa subsequently. Some of these non-canonical Hox functions are discussed in our review.

DOI: 10.1134/S0006297924060014

Keywords: homeodomain, ANTP, Hox genes, non-canonical functions of Hox genes, Metazoa, ParaHoxozoa,

Nephrozoa, neurogenesis, developmental autophagy, oogenesis, vertical signaling

Abbreviations: ANTP,Antennapedia; HOX,homeotic homeo-

box; GRN, gene regulatory network; PG, paralogue group;

Ubx,ultrabithorax.

* To whom correspondence should be addressed.

INTRODUCTION

The history of multicellular animal emergence is

tightly linked to the appearance of a new class of tran-

scription factors– Antennapedia (ANTP), members of

the superclass of homeodomain proteins [1]. Therapid

structural and functional evolution of the ANTP genes

led to the emergence of the most numerous and diverse

clade in the animal kingdom, now referred to as Para-

Hoxozoa [2]. This clade includes approximately 7mil-

lion species of bilaterians (Bilateria), about 10thousand

species of cnidarians (Cnidaria), and several species of

placozoans (Placozoa). The name of the clade indicates

an evolutionary boundary within Metazoa, which sep-

arates taxa with Hox/ParaHox genes from comb jellies

and sponges, which lack these genes [2-5].

Hox (homeotic homeobox) genes were the first

genes shown to be involved in development and evo-

lution [6]. Their discovery led to the emergence of a

new science – evolutionary developmental biology

(evo-devo) and made Hox genes the most stud-

ied group among all homeobox genes in animals.

The homeobox is a conserved region of the primary

sequence that encodes a DNA-binding motif of the

homeodomain, which is required for the Hox protein

to interact with enhancers of downstream target genes

[7, 8]. Hox genes are organized in a cluster, i.e., they

are physically linked. Traditionally, they are divid-

ed into 9 paralogous groups (PG1-8 and PG9/14). This

classification is based on the differences in Hox pro-

tein sequences and their relative positions in clusters.

The level of evolutionary conservation within a pa-

ralogous group (e.g., between lab (PG1) of the fruit fly

and Hox1 (PG1) of the amphioxus) is always higher

than outside of it (between lab (PG1) and pb (PG2) of

the fruit fly). Overall, the structural expansion of the

Hox cluster and the formation of the most paralogous

KULAKOVA et al.988

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

groups occurred before the appearance of the three

major Bilaterian clades [9, 10].

The main feature of the Hox cluster is the abili-

ty to exhibit a colinear expression. Colinearity refers

to the correspondence between the location of genes

on a chromosome and the order of their expression

along the anterior-posterior axis of the body [11, 12].

Thecloser is a Hox gene to the 3′-end of the cluster, the

closer to the anterior end of the embryo it will func-

tion. This is called spatial colinearity. Colinearity can

also be temporal, when genes are expressed sequential-

ly in time, starting from the 3′-end of the cluster [13].

The Hox gene cluster is the result of tandem cis-du-

plications of the ancestral sequence, starting from the

single proto-Hox gene that belonged to the NK family

[14, 15]. This event occurred before the sister branches

Bilateria and Cnidaria were formed because the Hox

genes belonging to the paralogous groups PG1, PG2,

and PG4/14 are already present in both these branch-

es [16]. These paralogous groups emerged as a result

of the diversification of cis-duplicates under two sce-

narios: neofunctionalization and subfunctionalization.

In the former case, a copy acquires a new function,

while in the latter, the ancestral function is shared

between two copies [17]. The cluster can be intact

(compact or relaxed) or contain rearrangements and

discontinuities (breakages) up to complete atomiza-

tion [18-21]. Integrity of the Hox cluster is essential for

maintaining temporal colinearity but is not crucial for

spatial colinearity [22].

Paradoxically, Hox genes are both highly con-

served and functionally flexible. Their functions are

universal during the establishment of the bilateral

body plan and species-specific at the level of local pat-

terns, such as the formation of bristles on the legs of

different Drosophila species [23]. This systemic prop-

erty is known as scalability. In the case of Hox genes,

it is manifested at both ontogenetic and phylogenetic

levels.

The staggering diversity of the bilaterian animals

results from the rapid evolution of developmental pro-

grams, which are simultaneously stable and flexible.

Bilaterian animals maintain a common body plan due

to the conserved members and parts of gene regula-

tory networks (GRNs), which start functioning shortly

after cleavage and are necessary for regionalization

and patterning. This is particularly evident in verte-

brates and other segmented animals, which undergo a

phylotypic period or “zootype” stage [24]. During this

period, representatives of a particular phylum (or sub-

phylum) are morphologically very similar to each other

(for example, all vertebrates at the pharyngula stage).

Athe molecular regulation level, the similarity is even

broader, as segmented animals from different phyla

(vertebrates, arthropods, annelids) exhibit orderly ex-

pression of Hox genes shortly before or during gastru-

lation [25-27]. The zootype concept is graphically rep-

resented as the hourglass model, where the constricted

waist corresponds to the onset of colinear transcription

of the Hox clusters along the anterior-posterior axis of

the body. Such expression is conceptually similar in all

segmented Bilateria, despite significant differences in

implementation mechanisms. Hox proteins determine

the fate of cells in broad spatial domains of the embryo

along the anterior-posterior axis of the body. During

this early period, their targets are genes of signaling

pathways and transcription factors, the sets of which

will qualitatively and quantitatively differ depending

on the Hox code. This difference ultimately will lead

to morphological and functional differences between

embryo regions.

This function is traditionally considered basic,

i.e., canonical. It is this function that is implied when

discussing the role of Hox genes in development, and

there are several reasons for this. Firstly, animals in

which we observe early colinear transcription of Hox

genes in broad spatial domains (i.e., regionalizing func-

tion) belong to the three superphyla– Deuterostomia,

Ecdysozoa, and Lophotrochozoa, which are grouped

into the clade Nephrozoa. The likelihood that such a

function of the Hox genes arose independently (conver-

gently) in these groups appears lower than the likeli-

hood of its direct inheritance from a common ancestor

of all Nephrozoa. Secondly, the sequential in-time ear-

ly activation of Hox cluster genes is always associated

with regionalization, and such a mode of activation

depends on the intactness or minimally damaged state

of the cluster [22, 28]. If the structure of the cluster

determines its regionalizing function and if a whole

cluster is apriori considered primary, it is logical to

assume that this function itself is inseparably linked to

the emergence of the Hox cluster.

The constricted waist in the hourglass model in-

dicates a lack of variability available for selection,

so it can be confidently stated that the action of Hox

genes determines the organization of the body plan, at

least in the segmented Bilateria. However, with this ap-

proach, questions about the early stages of system evo-

lution remain unsolved. The coordinated temporal and

spatial early colinear expression of Hox clusters looks

very complex. It is difficult to imagine the primary and

intermediate steps that formed this hyper network.

Moreover, if the canonical function of Hox genes is pri-

mary, then the last common ancestor of Nephrozoa is

a complex animal, not inferior in terms of organization

to the amphioxus, fruit fly, or Platynereis. This leads us

to the old paradox of irreducible complexity. Perhaps

among the multitude of functions of Hox genes, which

are not canonical ones, there are those that provide a

clue to the primary state of Hox regulation in the Para-

Hoxozoa lineage and its subsequent evolution. Some of

them are discussed in our review.

IRREDUCIBLE COMPLEXITY OF HOX GENE 989

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

HOX GENES AND NEUROGENESIS

The older the trait is, the more likely it is to be

found in the phylogenetically distant descendants of

the species that acquired this trait. The idea that the

ancestral function of Hox genes is patterning of the

nervous system was first suggested by Jordi Garcia

Fernandez [29], and not without reason. If we put the

canonical function of Hox genes out of consideration,

we are left with the most widespread and most dam-

age-resistant function– control of neurogenesis [30-32].

Control persists even in the animals that have aban-

doned early regionalizing function (leeches, appendic-

ularia, rotifers), or use it in isolation from the anteri-

or-posterior axis specification (mollusks) [19-21, 33-36].

It is noteworthy that in most of the studied mollusks,

the ganglia of the nervous system show signs of Hox

colinearity [21, 35, 36].

Modern experimental methods make it possible to

locally turn on or off selected genes at different stages

of development of model animals (nematodes, Drosoph-

ila, mouse). These experiments revealed an important

pattern. It turned out that the Hox genes do not simply

determine the cellular territories in which neuroblasts

are laid down. But they control pathways of their dif-

ferentiation and, more interestingly, establish terminal

specificity of the mature postmitotic neurons [32]. Ter-

minal specificity (neuronal terminal identity) brings

the neuron to a functional state. It begins to form neu-

rites, synthesize neuropeptides, proteins necessary

for the production of neurotransmitters, receptors to

them, and components of ion channels. All these and

many other changes in the postmitotic neurons occur

due to unrelated regulatory proteins called terminal

selectors. One such selector in the nematode Caenor-

habditis elegans, Unc-3 (an ortholog of EBF/Olf/Collier),

determines the terminal differentiation of the cholin-

ergic motor neurons. The Unc-3 protein binds directly

to the cis-regulatory sites of acetylcholine biosynthesis

genes, ion channels, and numerous other genes. It does

not work alone but in cooperation with various Hox

proteins that act as its cofactors. Different Hox proteins

(depending on the site of the body) determine differ-

ences in the number and length of neurites, synaptic

connections, and electrical activity of the motor neu-

rons. This general scheme is also valid for other types

of neurons (sensory, motor, and intermediate) with

other terminal selectors [37]. Itis important that the

Hox proteins are needed by the worm not only for the

correct tuning of the neuron at the time of its termi-

nal differentiation but also for its further work. It has

been shown that the Hox protein of C. elegans Lin-39

(PG 4/5) is necessary in adulthood to maintain the ter-

minal specificity of the motor neurons [38, 39].

C. elegans is a simply organized animal. An adult

worm (hermaphrodite) has only 302 mature neurons[40].

An adult Drosophila brain contains about 200,000 neu-

rons [41], but their differentiation, targeting, and num-

ber of synapses are determined by similar processes.

The fly neuroblasts acquire unique fates under the

action of Hox proteins [42], and, more surprisingly,

neuromuscular synapse formation is under the Hox

control [31, 43]. It is hypothesized that the assembly of

synaptic contact between a neuron and a muscle cell is

possible if they express the same Hox protein (or set of

Hox proteins). This is true for at least one model sys-

tem, where the Hox protein Dfd (PG4) directly turns

on the expression of ankyrin (Ankyrin2-XL; synaptic

protein) and turns off the expression of Con (an adhe-

sive protein that selectively works in neuromuscular

synapses of another type) in the motor neurons and

muscles they innervate [43].

Finally, a type of neurons (leucokinergic neurons)

has been described in Drosophila for which Hox pro-

teins from the BX-C complex [Ubx (ultrabithorax),

abd-A (abdominalA), Abd-B (abdominalB)] are direct

terminal selectors because they turn on (Ubx, abd-A)

and turn off (Abd-B) synthesis of the neuropeptide leu-

cokinin [44].

Without going into details, it should be noted that

mammals (mouse, human) have fundamentally similar

rules for establishing proneural territories, differenti-

ation of neurons, and their postmitotic settings under

the control of Hox genes [30]. Mammals have also been

shown to have Hox proteins that function as terminal

selectors of motor neurons [45] and are required in

adulthood. Hindbrain development is controlled by 24

Hox genes, and they continue to work in the mature

brains of adult mice, while forebrain development in

vertebrates is a Hox independent process [46].

This is all the more surprising that the Hox genes

from several anterior paralogical groups (PG1,3-5)

begin to be expressed in the postnatal neocortex and

thalamus of the mouse [46].

Thus, modern experimental data obtained with

different model animals lead us to the idea that the

ancestral function of Hox genes is the terminal dif-

ferentiation of neurons, most likely motor neurons.

Thishypothesis has a strong theoretical and evidence

base. Firstly, it has recently become known that the

homeobox-containing factors in general tend to trigger

and maintain neurogenic differentiation. The C. ele-

gans genome encodes 102 homeodomain-containing

proteins from different families that selectively and

combinatorially function as terminal selectors or their

partners in the mature neurons [47, 48]. Therefore, the

specification of neurons using Hox proteins is a special

case of a general principle.

Secondly, in the GRNs, direct linkage between

high-level regulatory genes and terminal differentia-

tion genes may indicate the ancestral state of the sys-

tem. Within the GRNs master genes, target genes and

KULAKOVA et al.990

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

intermediate regulatory genes are distinguished. These

intermediate genes form a complex architecture of the

developmental GRNs, and according to the hypothesis

of intercalary evolution, they are the result of evolu-

tionary intercalations (insertions) between the master

gene and its target. Such as in the case of the homeo-

box gene Pax6 and the light-sensitive transmembrane

protein rhodopsin [49-51]. The Hox genes are univer-

sally implicated in the establishment and maintenance

of the terminal neuronal specification in Protostomia

and Deuterostomia, allowing for the existence of a sim-

ple organized ancestor of all Nephrozoa, which used

Hox genes for the same purpose. Initially, the simple

GRNs of such ancestor gradually and independently be-

came more complex in different evolutionary lineages

due to the involvement of the new clade-specific genes

under the control of the Hox cluster. In the course of

evolution, heterochronies shifted the activity onset of

all participants to earlier developmental stages and

brought their expression to the canonical state. It is

possible that in the first stages of this evolutionary pro-

cess, Hox genes coordinated the formation of synapses

between the motor neurons and muscles. Therefore,

the general principle of colinear transcription of Hox

genes relates Drosophila and mammals at the level of

two germ layers, ectodermal and mesodermal. Impor-

tantly, the principle of intercalary evolution allows the

GRN growth by duplication of the master gene and sub-

functionalization of the descendant genes with partial

preservation of the ancestral function [51].

This attractive hypothesis has internal contradic-

tions. Firstly, the physical linkage of terminal selectors

is not necessary to specify neurons. Most homeobox

genes that create the neural code in nematodes are not

clustered [47, 48]. Secondly, the level of complexity of

the last common ancestor of all Nephrozoa remains

in question because its Hox cluster already consisted

of at least 7 or 8 genes– five anterior (PG1-5), one or

two middle (PG6/8), and one posterior (PG9/14) [9]. It is

known that functions of the genes from different paral-

ogous groups overlap significantly [52], which means

that the ancestral cluster formed very quickly before

its members began to differ greatly in the spectra of

their targets. If the quantitative information realized

by the Hox proteins was at some point more important

than the qualitative (paralog-specific) information, this

could have pushed the Hox cluster to a rapid structur-

al expansion with minimal divergence of participants.

However, the paralog-specific functions began to ap-

pear later. This explanation looks logical but raises the

following questions: why are the Hox proteins from

different paralogous groups qualitatively important

for the neuron specification, and why do these proteins

differ structurally, while some of their paralog-specif-

ic functions are conserved (common to Nephrozoa)?

It seems that selection drove the evolution of the Hox

cluster in several directions at once, and this could be

explained by the scenario when the neurogenic func-

tion was not the only one. Whether this is true or not,

we could figure it out by referring to basal taxa.

Outside the Nephrozoa group, expression of the

Hox genes has been studied less, but it is known that

small and disjointed clusters of Acoelomorpha (a sister

branch of Nephrozoa, previously classified as a clade

within flatworms) operate in the nervous, muscular,

and reproductive systems [53-56]. In a single study

on embryos [53], three Hox genes of Convolutriloba

longifissura (PG1, PG5, and PG9/14) colinearly turn on

in the proneural territories shortly after gastrulation.

Two of these three genes operate in parenchymatous

internal domains slightly later or simultaneously with

thisevent.

The cnidarian Hox genes have been extensively

studied [57-59], and expression of some of them can

be associated with the nervous system, e.g., Hox1 (PG1)

of Clytia hemisphaerica functions in statocysts and

Anthox1 (PG9-like) in the apical tuft of the Nematostella

vectensis planula larva. However, in comparison with

the diverse neural differentiation involving other ho-

meobox genes in cnidarians, this is a very modest

outcome [60, 61]. Surprisingly, many cnidarian neu-

rotransmitters (including acetylcholine) and enzymes

of their biogenesis are synthesized not in neurons but

in gastrodermal cells [62]. Analyzing expression data

is also difficult because direct correspondence of the

genes from Hox/ParaHox classes in cnidarians and bi-

laterians is not obvious due to the high divergence or

loss of orthologs [63]. However, there are no direct reg-

ulators of neurogenesis among the genes, which exact-

ly match the category of “PG1-like” or “PG2-like”, but

there are genes with broad expression domains at the

level of ectoderm and endomesoderm.

Thus, before the origin of the last common ances-

tor of Nephrozoa, at the level of Acoelomorpha, Hox

genes were already engaged in several different de-

velopmental programs. Their functions in cnidarians

are diverse and not associated with terminal neuronal

specification. The path from the common ancestor of

cnidarians and bilaterians to modern Nephrozoa was

accompanied by structural expansion of the Hox clus-

ter and complex rearrangements of the regulatory rela-

tionships between the ancient developmental programs

hidden to us. Some of these programs may have relied

on the shared paralog-nonspecific functions of Hox

genes, which are realized in isolation from the spa-

tial colinear transcription. We hypothesize that these

functions have remained in modern animals, and to

investigate them we need to look at the general biolog-

ical processes in which Hox genes are involved. There

are several examples where the paralog-nonspecific

function of Hox genes is realized particularly clearly.

Wewill discuss them in the following sections.

IRREDUCIBLE COMPLEXITY OF HOX GENE 991

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

HOX PROTEINS CONTROL DEVELOPMENTAL

TIMING THROUGH AUTOPHAGY

Autophagy is a cellular degradation process neces-

sary for maintaining cell homeostasis and renewing its

cytoplasmic components. It is highly conserved, and its

influence on various biological functions has been de-

scribed in a wide range of organisms, from plants and

yeast to humans [64]. For bilaterians, autophagy is an

important tool in early development, as it participates

in cellular differentiation and tissue remodeling [65,

66]. For instance, in the Drosophila larvae, autophagy

activity is very high in the fat body during the wander-

ing L3 (L3W) stage, when the larva is rapidly growing

and undergoing metamorphosis, but not in the younger

feeding L3 (L3F) stage. It has been shown that the tran-

sition from the L3F to L3W stage is controlled by ecdys-

one, and the main regulators of autophagy in this case

are Hox proteins, which suppress premature autopha-

gy at the L3F stage [67]. During normal development,

colocalization of Hox proteins from multiple paralo-

gous groups is observed in the fat body of L3F larvae,

but it is not colinear. Here, the Hox proteins suppress

the expression of the atg genes (18 genes), which are

responsible for autophagy. It has been shown that in-

activation of the individual Hox genes (Dfd, Scr, Ubx,

abd-A, AbdB) does not lead to premature induction of

autophagy. Only the simultaneous shutdown of all Hox

paralogs in the L3F larval experiment initiates this

process [67]. Conversely, prolonged expression of the

investigated Hox genes inhibits autophagy in the lar-

val fat body cells. Such animals enter the wandering

stage 6-7 days later than the controls, indicating that

the forced maintenance of the Hox gene expression de-

lays development in Drosophila.

Thus, in the larval fat body, the universal activi-

ty of Hox proteins carries temporal rather than spa-

tial information, regulating the onset of autophagy at

the required stage of development. It is worth noting

that in the culture of mammalian fibroblasts, HoxB8

and HoxA9 also inhibit autophagy. The same proteins

exhibit a similar effect on the Drosophila larvae after

transgenesis [67]. These preliminary studies do not rule

out the paralog-nonspecific involvement of Hox genes

in autophagy control in the last common ancestor of

insects and vertebrates but require additional analysis

across a wide range of models.

WHY DO HOX GENES WORK

BEFORE DIFFERENTIATION BEGINS?

In multicellular animals, Hox genes are not ex-

pressed in totipotent and pluripotent cells because this

expression induces differentiation. In mammalian em-

bryonic stem cells, Hox loci have an ambivalent epigen-

etic status. Their histone code contains both repressive

and permissive tags [68]. These cells do not express

Hox genes, but expression can start rapidly in case of

additional permissive signals, which will lead the cells

to the beginning of the differentiation path.

Despite the blockage of Hox genes activity in toti-

potent cells, their maternal transcripts have been found

in oocytes of mammals (mouse, cow, and human [69],

amphibians (Xenopus laevis [70]), annelids (Platynereis

dumerilii [71]), myriapods (Strigamia maritima [72];

Trigo niulus corallinus [73], hymenopterans (ants of the

tribe Camponotini [74]), decapods (Macrobrachium olfer-

sii [75]), and even in oocytes of hydroid polyp Clitya

hemisphaerica [58]. In all animals, except Xenopus lae-

vis, the genes from several paralogous groups are ex-

pressed in oocytes.

The structure of oocyte transcripts of Hox genes

may provide clues about their functions. The exam-

ple of centipede Strigamia maritime (Chilopoda) [72]

showed that the maternal RNAs of Hox genes are

polyadenylated, but some of them do not contain an

open reading frame. Possibly, part of the maternal

RNAs of the centipede belongs to the class of regulatory

(protein-noncoding) RNAs. On the other hand, the Hox

gene transcripts in mammalian oocytes are deadenylat-

ed [69]. This is reasonable if such maternal transcripts

are required for later developmental stages and stored

in a stable (non-translated) form [76]. In addition, the

HOXB9 protein was detected in the nuclei of oocytes

and the cells of early mammalian (mouse, cow) embry-

os [77], and the Ubx and AbdA proteins were found in

ant oocytes [74].

The function of Hox genes in oocytes differs from

the canonical one since transcripts of different paral-

ogous groups are localized in a single cell. The wide-

spread nature of this phenomenon suggests that it is

not random. So far, there are no successful experiments

unambiguously indicating the functions of the Hox

gene transcripts in oocytes, but several hypotheses can

be put forward.

Most of the maternal transcripts of the Hox genes

may not be translated. They may represent an element

of epigenetic tuning of the zygotic genome. It is known

that the non-coding RNAs often function as scaffolds

for the assembly of chromatin remodeling proteins,

targeting them to subordinate loci. This function has

been described for the regulatory RNAs that are read

from the vertebrate Hox clusters [78] and it is consis-

tent with the presence of transcripts without an open

reading frame in the Strigamia oocytes. The importance

of the non-coding RNAs in the first cleavage divisions

was shown in the mouse embryos [79].

On the other hand, the mRNAs of Hox genes

can be translated. For example, the HoxB9 protein

is present in the nuclei of oocytes (both mature and

immature) and the cell nuclei of early embryos [77].

KULAKOVA et al.992

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

This doesnotexclude an early function of Hox genes

directed to oogenesis. It is known that in the mouse

several transcripts of Hox genes and their cofactors are

already present at the stage of the growing oocyte [80].

Interestingly, the homeodomain protein Nobox from the

family of the same name, which is close to both the Antp

and PRD classes [81], is present in the mouse oocytes

and regulates the functioning of the genes important

for oogenesis [82]. It cannot be ruled out that the oocyte

RNAs and proteins of the Hox genes are required for the

transcriptional control of the early zygotic genes [83].

The Hox proteins can be not only transcriptional

regulators, but also regulators of the cell cycle, RNA

splicing, DNA replication and repair [84-86]. For exam-

ple, some Hox proteins have recognition sites for the

serine/threonine kinase ATM [87], whose role in the

DNA double-strand break repair is widely known [88].

At least one vertebrate Hox protein, HoxB7, has been

experimentally shown to be involved in this process

[85]. In the mammary epithelial cells, HoxB7 increas-

es the probability of nonhomologous DNA end joining

by binding to the complex of Ku70/80 heterodimers

(proteins that recognize double-stranded breaks) and

DNA protein kinase [85]. HoxB7 expression stimulates

the DNA-protein kinase activity, which correlates with

the repair efficiency, whereas this effect is lost when

HoxB7 is knocked out. In addition, Hox proteins pro-

mote the assembly of pre-replicative complexes. For

example, HoxD13 and HoxC13 in different cell cultures

interact through homeodomains with the proteins

ofORC and Cdc6 pre-replicative complexes [86, 89, 90].

Although the mentioned functions of Hox proteins were

not described in embryogenesis, we assume that they

can still participate in the first stages of the develop-

ment of multicellular animals. In the early develop-

ment, synchronous divisions of blastomeres occur with

a minimum interval (there are no G1 and G2 phases of

the cycle), therefore, a very precise adjustment of the

molecular machinery of the oocyte is necessary for the

successful completion of the cleavage stage. In this case,

Hox proteins in oocytes may accelerate the assembly of

pre-replicative complexes and enhance DNA repair to

maintain the integrity of the embryo’s genetic material.

Undoubtedly, the role of Hox genes in oogenesis

will not be clarified without functional tests on a wide

range of models. It cannot be ruled out that the oocyte

RNAs and proteins of the Hox genes could be a “tran-

scriptional noise” or by-products of the previous stages

of oogenesis [91].

DOSE-DEPENDENT FUNCTIONS

OF HOX PROTEINS

There are dose-dependent functions of Hox pro-

teins when their concentration determines the mor-

phology of the anlage [92]. In mammals, Hox proteins

specify vertebral morphology in a dose-dependent

manner [93] and set the number and length of dig-

its [94]. In both cases, a gradual decrease in the dose

of Hox proteins increases intensity of morphological

changes. For instance, when the dose of any protein

from the posterior paralogs of HoxA and HoxD clus-

ters (Hoxd11, Hoxd12, Hoxa13, Hoxd13) is gradually

decreased, the digits shorten linearly and paralog-

independently depending on the proportion of mutant

alleles of the Hox genes [94].

In invertebrate animals, the most obvious exam-

ple of a dose-dependent function of Hox proteins is the

regulation of wing morphology. This function has been

described in insects from different clades and, appar-

ently, is universal for the diversification of the wing

shape and size in the second (T2) and third (T3) thorac-

ic segments [92]. It involves the Antp and Ubx proteins

in its realization. In the wild-type Drosophila, the Antp

protein is present only in T2 and the Ubx protein in

T3, with a lower concentration of Antp in T2 than of

Ubx in T3 (Fig.1a) [95]. In the Ubx

–/–

mutants, a pair

of wings is formed at T3 instead of halteres (Fig.1b).

If the dose of Antp in T3 is increased to the level of

Ubx in such mutants, the normal phenotype is restored

(halteres are formed on T3 (Fig.1c) [95]. On the con-

trary, when the Ubx dose is decreased, wings grow in-

stead of halteres at T3 (Fig.1d) [95]. Likewise, when

the Antp dose is increased in T2, halteres grow instead

ofwings (Fig.1e)[95].

HOX FACTORS CAN BE SECRETED BY CELLS

The difficulty faced by a researcher who decides to

elucidate the ancestral function of the ANTP-class ho-

meobox genes is related to the fact that such function

was not originally a single function, at least at the level

of Metazoa. This follows from the arrangement of mul-

titasking ANTP-class homeodomain proteins (Fig. 2).

Atthe end of the last century, it was discovered that a

synthetic homeodomain protein of 60 amino acid res-

idues, repeating the sequence of the Drosophila Antp

homeodomain, can penetrate the membranes of rat

nerve cells without the mediation of any receptors.

After penetration, it is transported into the nucleus

and increases the level of differentiation of the re-

cipient cells [96]. Later it was found out that the nat-

ural homeodomain proteins Emx1, Emx2, Engrailed-2

(En2), Hoxa5, Hoxb4, Hoxc8, Knotted1, Otx2, Pax6, and

Vax1 are present in the cells that do not express their

mRNA[97].

Apparently, most of the homeodomain proteins, in-

cluding those beyond the boundaries of the ANTP class

[97-100], have the ability for intercellular transport like

a signaling or morphogen molecule. The mechanism

IRREDUCIBLE COMPLEXITY OF HOX GENE 993

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

Fig. 1. Effect of Antp and Ubx protein dose on the Drosophila phenotype. a)“Wild-type” (WT); b)Ubx mutant –/–; c)Antp dosage

increase in T3 to the Ubx level restores normal phenotype; d)decrease of Ubx dosage in T3 to the Antp level in T2 leads to wing

formation; e) Antp dosage increase in T2 to the Ubx level in T3 leads to halteres formation; f) hypothetical phenotype that

could be modeled on Drosophila. It is characteristic of insects from the order Strepsiptera. The illustration is based on the data

ofPauletal. (2021) and Merabet and Carnesecchi (2024) [92, 95].

bywhich this is realized is not yet fully understood.

Itis known that secretion and internalization depend

on two overlapping motifs localized in the most con-

served regions of the homeodomain [100]. In addition,

secretion depends on individual hydrophobic amino

acids outside of the homeodomain (Fig.2) [97]. Itseems

that homeodomain proteins can enter any cell type

through macropinocytosis, but the efficiency of the

process depends on the structure of the glycocalyx of

the receiving cell [97, 101].

In the impressive study conducted in 2019 [97], 162

human homeodomain proteins from different classes

were tested for their secretion and transfer abilities,

and the test was performed simultaneously on three

different cell cultures (secretion – HEK 293T, GT1-7,

and MDCK; internalization – HeLa). It was shown

that secretion efficiency strongly depends on the cell

type and characteristics of the primary sequence of

the homeodomain proteins themselves. For example,

the proteins EN2, HOXC8, PAX6, and VAX1 were secreted

KULAKOVA et al.994

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

Fig. 2. Structural motifs of proteins from the ANTP class and their participation in the realization of the Hox protein functions.

The scheme describes the functional significance of the main domains of Hox factors. The homeodomain, the most conserved

region of ANTP-class proteins, is required for Hox factors to realize both canonical (transcription) and non-canonical functions.

In addition to the homeodomain, a high degree of conservatism is also characteristic of the short hexapeptide motif by which

Hox proteins interact with cofactors [102]. Surprisingly, the transport capacity of the Hox factors is also mediated by these do-

mains: the homeodomain contains secretion and internalizing (penetratin) motifs, and the conserved tryptophan residue in

the hexapeptide is required for Hox proteins to be exported from the nucleus [103]. For some Hox factors, it has been shown

that their secretion could depend on individual hydrophobic amino acids localized at less conserved sites on the protein [97].

Colored letters in the sequences indicate conserved positions, colored boxes surround motifs, non-canonical functions are in

bold; Mm,Mus musculus; Dm,Drosophila melanogaster; aa,amino acid residues.

in all three cell cultures, while HOXA5 and OTX2 were

secreted in only two. Ten proteins, including only

one Hox (HOXA10), were not secreted at all, which,

however, does not exclude this possibility in other

cell types. All tested proteins (including non-secreted

ones) turned out to be capable of internalization. This

means that secretion of the homeodomain proteins is

the crucial stage of transduction at which cells control

this process.

Homeodomain proteins can diffuse over two or

three-cell diameters like paracrine signaling factors,

but there are examples of their extensive diffusion.

In the mouse, the Otx2 protein is synthesized in the

vascular plexus, secreted into the cerebrospinal fluid,

and accumulates throughout the cerebral cortex [104].

Importantly, the endogenous Hox proteins prefer-

entially function as transcription factors. Immediately

after translation, they are taken up by karyopherins,

which recognize the nuclear localization signal, and

transport Hox proteins to the nucleus. The exogenous

Hox proteins exhibit a broader range of functions.

Ithas been shown that the secreted mouse Otx2 moves

into mitochondria, where it binds to mitochondrial ATP

synthases and enhances ATP synthesis [105]. Earlier

it was reported that the exogenous En2 in Xenopus

accumulates in the growth cones of neurons, controls

their axonal targeting, and indirectly enhances trans-

lation [106].

This does not exclude penetration of the exoge-

nous proteins into the nuclei, where they trigger tran-

scription of their mRNAs and other specific targets.

A classic example of this type is vertical signaling

during gastrulation in Xenopus. It was shown that the

Hox proteins from presomitic mesoderm sequentially

switch on the expression of their own Hox genes in the

gastrula neuroectoderm, i.e., copying of positional in-

formation from one germ layer to another occurs in

this case [104]. Direct exchange of transcription factors

between the cell layers coordinates the operation of de-

velopmental programs without the mediation of mor-

phogens and signaling cascades.

Remarkably, the signaling and regulatory func-

tions in general cases are mediated by the same evolu-

tionarily ancient and conserved motif– homeodomain.

The ability to solve two tasks with a single tool could

have been used to coordinate growth and development

by the first Metazoa, even before the emergence of

modern relationships between the long-range signaling

ligands, their messengers, and targets [104].

HOW DID NON-CANONICAL FUNCTIONS

BRING HOX CLUSTER ACTIVITY

TO A CANONICAL STATE?

The coherent operation of Hox genes required for

canonical function remains a great evolutionary mys-

tery because it is a multi-event process. It consists of:

– epigenetic tuning of Hox loci, including through

the regulatory RNAs encoded in the Hox clusters;

– establishment of the topology-associated do-

mains (TADs), which are stabilized according

tothe position of cells along the anteroposterior

axis;

– responses to multidirectional signals from mor-

phogens (retinoic acid, Wnt, Fgf, Bmp) in three-

dimensional coordinates of the embryo;

– coordinated expression in cells of different germ

layers due to the mechanism unique for home-

odomain proteins – vertical signaling;

IRREDUCIBLE COMPLEXITY OF HOX GENE 995

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

Fig. 3. Control of Hox cluster transcription. a)Some control mechanisms of the Hox cluster in bilateral animals. 1)GCR, Global

controlling element. Regulatory elements of this type have been described in vertebrates; 2)local cis-regulatory modules (in-

dividual and shared), site-specific transcription factors, and microRNAs (miR) encoded in Hox cluster sequences. These modes

ofregulation are present in protostomes and deuterostomes; 3)vertical signaling has been described in vertebrates. b)Non-ca-

nonical functions of Hox genes that could be realized by separate elements of the controls; c)Simplified scheme of Hox clus-

ter regulation during realization of the canonical function of axial patterning(d). Not all mechanisms are represented in the

scheme, and not all of those represented are universal. GCR and vertical signaling are not found in the model arthropods. Spiral

animals are generally poorly studied at the level of Hox cluster activation mechanisms, but among them, there are organisms

with temporal colinearity and early mesodermal transcription. RA (Retinoic Acid), FGF, and WNT are gradients of morphogens.

– response to individual signals from upstream

regulatory proteins, which can turn on/off indi-

vidual genes;

– reciprocal interactions of Hox genes (posterior

suppression and beyond).

Some of the regulatory mechanisms are shown

in Fig.3. How could this very complex picture emerge

from non-canonical functions? Perhaps the different

modes of Hox cluster regulation needed to perform

separate tasks were co-opted into a new program.

It seems intuitively correct to assume that this new

program is gastrulation in its “bilateral version.” This

seems to be suggested by the temporal conjugation of

the Hox cluster activation and gastrulation processes in

deuterostomes. Moreover, there is evidence for pre-ad-

aptation– two Hox genes of the cnidarian Nematostella

vectensis (anterior NvAx6 and mid-posterior NvAx1)

are important for gastrulation and specification of the

oral-aboral axis. Their expression sites label the oral

and aboral poles, and the morpholino knockdown sup-

presses gastrulation [107]. However, these genes do not

form a cluster and operate in different germ layers.

InDrosophila, the axial pattern of Hox genes is estab-

lished before the onset of gastrulation, and there are no

studies to reliably confirm or refute the involvement of

the Hox genes in arthropods in gastrulation. Insome

spiral animals (annelids, brachiopods, mollusks), early

activation of the Hox genes coincides with the onset

or continuation of gastrulation, but functional tests are

still lacking [26, 108, 109].

With a high degree of confidence, it can be stat-

ed that the last common ancestor of bilateral animals

KULAKOVA et al.996

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

hada single Hox cluster. This cluster could have been

used for different tasks, such as the specification of mo-

tor neurons, the establishment of neuromuscular junc-

tions, and a function requiring quantitative changes in

transcripts in response to a stimulus (probably a mor-

phogen concentration). Such an ancient function may

have been related to autophagy, control of prolifera-

tion, or gametogenesis. Importantly, this function was

enabled by the gradient distribution of Hox proteins

along the axis and kept the Hox genes in the cluster.

The Nephrozoa ancestor could use the whole cluster

or some of its genes to control gastrulation, but this is

not the case in the lineage of modern Acoelomorpha

(sister branch of Nephrozoa) because all Hox genes of

Convolutriloba are turned on after gastrulation [53].

Different ways of controlling the same cluster

could lead to errors in its regulation, and some of the

resulting aberrant variants were preserved by natural

selection. Two important events could have occurred in

the evolution of bilateral animals from the Nephrozoa

lineage. Firstly, several mechanisms controlling the

Hox cluster could have united to control a single mor-

phogenetic program. Secondly, there might have been

aheterochronic shift in the activation of this program

towards earlier development. The least catastrophic

variant suggests a series of heterochronic shifts of co-

linear expression of the Hox genes in internal, meso-

dermal in origin, structures up to the gastrula stage.

Then, through the vertical signaling, the Hox genes

began to turn on colinearly in the adjacent ectoderm

(future neuroectoderm), providing animals with a

new powerful tool for controlling early development.

Thistool could be easily scaled by the gradients of mor-

phogens, it coordinated the development of ectodermal

and mesodermal tissues, and it was evolutionarily plas-

tic due to many controls coming from the older pro-

grams. Perhaps it was this new molecular mechanism

that “detonated” and triggered the “Cambrian Explo-

sion” because of its ability to rapidly alter early devel-

opment.

CONCLUSION

The diversity of non-canonical functions of Hox

genes is determined by the structure of the home-

odomain protein itself, which can work not only as a

transcription factor, but also as a regulator of general

biological processes, such as DNA repair, replication,

translation, and RNA splicing.

The “hourglass” model, while illustrative, leaves

non-canonical functions of the Hox genes invisible.

According to the inverse hourglass model, which is

valid for Metazoa as a whole [110], there is a fundamen-

tal similarity in gene functioning at the earliest stages

of development (pluripotent state of cells, cleavage)

and at later stages (differentiation, organogenesis).

However, animals from different phyla will vary sig-

nificantly in the ensembles of regulatory genes and

nature of their involvement in morphogenesis in the

middle of development, just between cleavage and

committed differentiation [110]. These are the differ-

ences that define fundamental distinction between

the phyla within Metazoa. In other words, it is possi-

ble to identify distinct sets of signaling pathways and

transcription factors that interact during the estab-

lishment of Metazoa organization plans. Their specific

combination defines the appearance of each phylum.

Itturned out that homeobox genes in general and Hox

genes in particular do not fall into the category of such

“phylum- specific” regulators because their functions

are broader and more conserved during the divergent

period of development.

We assume that the proto-Hox gene initially pos-

sessed a wide repertoire of functions, some of which

relied on the signaling nature of its protein. Animals

from the ParaHoxozoa branch turned out to be heirs

of this regulatory complexity. They further enhanced it

through cooperation between developmental programs

that used different functional capabilities of Hox pro-

teins. These programs emerged at various stages of

evolution, and their traces are preserved in modern

animals in the form of distinct paralog-nonspecific and

dose-dependent functions.

Acknowledgments. In the study, Geneious®

2023.2.1 software was used for sequence analysis, with

access provided by the research resource center “Chro-

mas” of Saint Petersburg State University.

Contributions. M.A.K. developed the study con-

cept, supervised the study, and prepared and edited

the manuscript; G.P.M. prepared and edited the manu-

script; L.O.P. prepared and edited the manuscript.

Funding. This work was financially supported by

the Russian Science Foundation (project no.23-24-00426).

Ethics declarations. This work does not contain

any studies involving human and animal subjects.

Theauthors of this work declare that they have nocon-

flicts of interest.

Open access. This article is licensed under a Cre-

ative Commons Attribution4.0 International License,

which permits use, sharing, adaptation, distribution,

and reproduction in any medium or format, as long

as you give appropriate credit to the original author(s)

and the source, provide a link to the Creative Commons

license, and indicate if changes were made. Theimages

or other third-party material in this article are includ-

ed in the article’s Creative Commons license, unless

indicated otherwise in a credit line to the material.

Ifmaterial is not included in the article’s Creative Com-

mons license and your intended use is not permitted

by statutory regulation or exceeds the permitted use,

IRREDUCIBLE COMPLEXITY OF HOX GENE 997

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

you will need to obtain permission directly from the

https://creativecommons.org/licenses/by/4.0/.

REFERENCES

1. Degnan, B. M., Vervoort, M., Larroux, C., and Rich-

ards, G. S. (2009) Early evolution of metazoan tran-

scription factors, Curr. Opin. Genet. Dev., 19, 591-599,

doi:10.1016/j.gde.2009.09.008.

2. Ryan, J.F., Pang,K., Mullikin, J.C., Martindale, M.Q.,

and Baxevanis, A.D. (2010) The homeodomain com-

plement of the ctenophore Mnemiopsis leidyi suggests

that Ctenophora and Porifera diverged prior to the

ParaHoxozoa, Evodevo, 1, 9, doi:10.1186/2041-9139-1-9.

3. Pang,K., and Martindale, M.Q. (2008) Developmen-

tal expression of homeobox genes in the ctenophore

Mnemiopsis leidyi, Dev. Genes Evol., 218, 307-319,

doi:10.1007/s00427-008-0222-3.

4. Srivastava, M., Begovic, E., Chapman, J., Putnam,

N.H., Hellsten,U., Kawashima,T., Kuo,A., Mitros,T.,

Salamov,A., and Carpenter, M.L. (2008) The Trichop-

lax genome and the nature of placozoans, Nature,

454, 955-960, doi:10.1038/nature07191.

5. Pastrana, C.C., DeBiasse, M.B., and Ryan, J.F. (2019)

Sponges lack ParaHox genes, Genome Biol. Evol., 11,

1250-1257, doi:10.1093/gbe/evz052.

6. Lewis, E. B. (1978) A gene complex controlling seg-

mentation in Drosophila, Nature, 276, 565-570,

doi:10.1038/276565a0.

7. Korchagina, N. M., Bakalenko, N. I., and Kulakova,

M.A. (2010) Hox-cluster and evolution of morphogen-

eses, Russ.J. Dev. Biol., 41, 353-363.

8. Hubert, K. A., and Wellik, D. M. (2023) Hox genes

in development and beyond, Development, 150,

dev192476, doi:10.1242/dev.192476.

9. De Rosa, R., Grenier, J. K., Andreeva, T., Cook, C. E.,

Adoutte,A., Akam,M., Carroll, S.B., and Balavoine,G.

(1999) Hox genes in brachiopods and priapulids

and protostome evolution, Nature, 399, 772-776,

doi:10.1038/21631.

10. Balavoine, G., de Rosa, R., and Adoutte, A. (2002)

Hox clusters and bilaterian phylogeny, Mol. Phy-

logenet. Evol., 24, 366-373, doi: 10.1016/s1055-7903

(02)00237-3.

11. Dressler, G. R., and Gruss,P. (1989) Anterior bound-

aries of Hox gene expression in mesoderm-derived

structures correlate with the linear gene order

along the chromosome, Differentiation, 41, 193-201,

doi:10.1111/j.1432-0436.1989.tb00747.x.

12. Duboule, D., and Dolle, P. (1989) The structural and

functional organization of the murine HOX gene

family resembles that of Drosophila homeotic genes,

EMBOJ., 8, 1497-1505, doi:10.1002/j.1460-2075.1989.

tb03534.x.

13. Duboule, D. (1994) Temporal colinearity and the

phylotypic progression: a basis for the stability of a

vertebrate Bauplan and the evolution of morpholo-

gies through heterochrony, Dev. Suppl., doi:10.1242/

dev.1994.Supplement.135.

14. Larroux, C., Fahey, B., Degnan, S. M., Adamski, M.,

Rokhsar, D.S., and Degnan, B. M. (2007) The NK ho-

meobox gene cluster predates the origin of Hox genes,

Curr. Biol., 17, 706-710, doi:10.1016/j.cub.2007.03.008.

15. Copley, R. R. (2023) The ancestry of Antennapedia-

like homeobox genes, bioRxiv, doi: 10.1101/2023.

03.14.532566.

16. DuBuc, T. Q., Ryan, J. F., Shinzato, C., Satoh, N., and

Martindale, M. Q. (2012) Coral comparative genom-

ics reveal expanded Hox cluster in the cnidarian-

bilaterian ancestor, Integr. Comp. Biol., 52, 835-841,

doi:10.1093/icb/ics098.

17. Lynch, V. J., Roth, J. J., and Wagner, G. P. (2006)

Adaptive evolution of Hox-gene homeodomains

after cluster duplications, BMC Evol. Biol., 6, 86,

doi:10.1186/1471-2148-6-86.

18. Duboule, D. (2007) The rise and fall of Hox gene

clusters, Development, 134, 2549-2560, doi: 10.1242/

dev.001065.

19. Seo, H. C., Edvardsen, R. B., Maeland, A. D., Bjord-

al, M., Jensen, M. F., Hansen, A., Flaat, M., Weissen-

bach, J., Lehrach,H., Wincker, P., Reinhardt, R., and

Chourrout,D. (2004) Hox cluster disintegration with

persistent anteroposterior order of expression in

Oikopleura dioica, Nature, 431, 67-71, doi: 10.1038/

nature02709.

20. Fröbius, A. C., and Funch, P. (2017) Rotiferan Hox

genes give new insights into the evolution of meta-

zoan bodyplans, Nat. Commun., 8, 9, doi: 10.1038/

s41467-017-00020-w.

21. Lee, P. N., Callaerts, P., De Couet, H. G., and Mar-

tindale, M. Q. (2003) Cephalopod Hox genes and

the origin of morphological novelties, Nature, 424,

1061-1065, doi:10.1038/nature01872.

22. Monteiro, A. S., and Ferrier, D. E. (2006) Hox genes

are not always Colinear, Int. J. Biol. Sci., 2, 95-103,

doi:10.7150/ijbs.2.95.

23. Stern, D. L. (1998) A role of Ultrabithorax in mor-

phological differences between Drosophila species,

Nature, 396, 463-466, doi:10.1038/24863.

24. Slack, J. M., Holland, P.W., and Graham, C.F. (1993)

The zootype and the phylotypic stage, Nature, 361,

490-492, doi:10.1038/361490a0.

25. Lemaire, L., and Kessel, M. (1997) Gastrulation and

homeobox genes in chick embryos, Mech. Dev.,

67, 3-16, doi:10.1016/s0925-4773(97)00102-0.

26. Kulakova,M., Bakalenko,N., Novikova,E., Cook, C.E.,

Eliseeva,E., Steinmetz, P.R., Kostyuchenko, R.P., Don-

dua,A., Arendt,D., Akam,M., and Andreeva,T. (2007)

Hox gene expression in larval development of the

polychaetes Nereis virens and Platynereis dumerilii

KULAKOVA et al.998

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

(Annelida, Lophotrochozoa), Dev. Genes Evol., 217,

39-54, doi:10.1007/s00427-006-0119-y.

27. Michaut,L., Jansen, H.J., Bardine,N., Durston, A.J.,

and Gehring, W. J. (2011) Analyzing the function

of a hox gene: an evolutionary approach, Dev.

Growth Differ., 53, 982-993, doi:10.1111/j.1440-169X.

2011.01307.x.

28. Fröbius, A. C., Matus, D. Q., and Seaver, E. C. (2008)

Genomic organization and expression demonstrate

spatial and temporal Hox gene colinearity in the

lophotrochozoan Capitella sp. I, PLoS One, 3, e4004,

doi:10.1371/journal.pone.0004004.

29. Garcia-Fernàndez, J. (2005) The genesis and evolu-

tion of homeobox gene clusters, Nat. Rev. Genet., 6,

881-892, doi:10.1038/nrg1723.

30. Philippidou, P., and Dasen, J. S. (2013) Hox genes:

choreographers in neural development, architects of

circuit organization, Neuron, 80, 12-34, doi:10.1016/

j.neuron.2013.09.020.

31. Joshi, R., Sipani, R., and Bakshi, A. (2021) Roles of

Drosophila Hox genes in the assembly of neuromus-

cular networks and behavior, Front. Cell Dev. Biol.,

9, 786993, doi:10.3389/fcell.2021.786993.

32. Feng,W., Li,Y., and Kratsios,P. (2021) Emerging roles

for hox proteins in the last steps of neuronal devel-

opment in worms, flies, and mice, Front. Neurosci.,

15, 801791, doi:10.3389/fnins.2021.801791.

33. Kourakis, M.J., Master, V.A., Lokhorst, D.K., Nardel-

li-Haefliger, D., Wedeen, C.J., Martindale, M.Q., and

Shankland,M. (1997) Conserved anterior boundaries

of Hox gene expression in the central nervous sys-

tem of the leech Helobdella, Dev. Biol., 190, 284-300,

doi:10.1006/dbio.1997.8689.

34. Samadi,L., and Steiner,G. (2010) Expression of Hox

genes during the larval development of the snail,

Gibbula varia (L.)-further evidence of non-colin-

earity in molluscs, Dev. Genes Evol., 220, 161-172,

doi:10.1007/s00427-010-0338-0.

35. Hinman, V. F., O’Brien, E. K., Richards, G. S., and

Degnan, B. M. (2003) Expression of anterior Hox

genes during larval development of the gastropod

Haliotis asinina, Evol. Dev., 5, 508-521, doi: 10.1046/

j.1525-142x.2003.03056.x.

36. Barrera Grijalba, C. C., Rodríguez Monje, S. V.,

Gestal,C., and Wollesen,T. (2023) Octopod Hox genes

and cephalopod plesiomorphies, Sci. Rep., 13, 15492,

doi:10.1038/s41598-023-42435-0.

37. Kratsios,P., Kerk, S.Y., Catela,C., Liang,J., Vidal,B.,

Bayer, E. A., Feng, W., De La Cruz, E. D., Croci, L.,

Consalez, G.G., Mizumoto, K., and Hobert,O. (2017)

An intersectional gene regulatory strategy defines

subclass diversity of C.elegans motor neurons, Elife,

6, e25751, doi:10.7554/eLife.25751.

38. Feng, W., Li, Y., Dao, P., Aburas, J., Islam, P., El-

baz, B., Kolarzyk, A., Brown, A. E., and Kratsios, P.

(2020) A terminal selector prevents a Hox tran-

scriptional switch to safeguard motor neuron iden-

tity throughout life, Elife, 9, e50065, doi: 10.7554/

eLife.50065.

39. Li,Y., Osuma,A., Correa,E., Okebalama, M.A., Dao,P.,

Gaylord, O., Aburas, J., Islam, P., Brown, A. E., and

Kratsios, P. (2020) Establishment and maintenance

of motor neuron identity via temporal modular-

ity in terminal selector function, Elife, 9, e59464,

doi:10.7554/eLife.59464.

40. White, J.G., Southgate,E., Thomson, J.N., and Bren-

ner,S. (1986) The structure of the nervous system of

the nematode Caenorhabditis elegans, Philos. Trans.

R Soc. Lond. B Biol. Sci., 314, 1-340, doi: 10.1098/

rstb.1986.0056.

41. Raji, J. I., and Potter, C. J. (2021) The number of

neurons in Drosophila and mosquito brains, PLoS

One, 16, e0250381, doi: 10.1371/journal.pone.

0250381.

42. Estacio-Gómez, A., and Díaz-Benjumea, F. J. (2014)

Roles of Hox genes in the patterning of the central

nervous system of Drosophila, Fly (Austin), 8, 26-32,

doi:10.4161/fly.27424.

43. Friedrich, J., Sorge, S., Bujupi, F., Eichenlaub, M. P.,

Schulz, N. G., Wittbrodt, J., and Lohmann, I. (2016)

Hox function is required for the development

and maintenance of the Drosophila feeding mo-

tor unit, Cell Rep., 14, 850-860, doi: 10.1016/j.celrep.

2015.12.077.

44. Estacio-Gómez, A., Moris-Sanz, M., Schäfer, A. K.,

Perea,D., Herrero,P., and Díaz-Benjumea, F.J. (2013)

Bithorax-complex genes sculpt the pattern of leucoki-

nergic neurons in the Drosophila central nervous

system, Development, 140, 2139-2148, doi: 10.1242/

dev.090423.

45. Catela,C., Chen,Y., Weng,Y., Wen,K., and Kratsios,P.

(2022) Control of spinal motor neuron terminal dif-

ferentiation through sustained Hoxc8 gene activity,

Elife, 11, e70766, doi:10.7554/eLife.70766.

46. Hutlet, B., Theys, N., Coste, C., Ahn, M. T., Doshish-

ti-Agolli, K., Lizen, B., and Gofflot, F. (2016) System-

atic expression analysis of Hox genes at adulthood

reveals novel patterns in the central nervous system,

Brain Struct. Funct., 221, 1223-1243, doi: 10.1007/

s00429-014-0965-8.

47. Reilly, M. B., Cros,C., Varol, E., Yemini, E., and Hob-

ert, O. (2020) Unique homeobox codes delineate all

the neuron classes of C.elegans, Nature, 584, 595-601,

doi:10.1038/s41586-020-2618-9.

48. Hobert, O. (2021) Homeobox genes and the specifi-

cation of neuronal identity, Nat. Rev. Neurosci., 22,

627-636, doi:10.1038/s41583-021-00497-x.

49. Sheng,G., Thouvenot,E., Schmucker,D., Wilson, D.S.,

and Desplan, C. (1997) Direct regulation of rhodop-

sin 1 by Pax-6/eyeless in Drosophila: evidence for a

conserved function in photoreceptors, Genes Dev.,

11, 1122-1131, doi:10.1101/gad.11.9.1122.

IRREDUCIBLE COMPLEXITY OF HOX GENE 999

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

50. Gehring, W. J., and Ikeo, K. (1999) Pax 6: mastering

eye morphogenesis and eye evolution, Trends Genet.,

15, 371-377, doi:10.1016/s0168-9525(99)01776-x.

51. Gehring, W.J. (2005) New perspectives on eye devel-

opment and the evolution of eyes and photorecep-

tors, J.Hered., 96, 171-184, doi:10.1093/jhered/esi027.

52. Merabet,S., and Mann, R.S. (2016) To be specific or

not: the critical relationship between Hox and TALE

proteins, Trends Genet., 32, 334-347, doi:10.1016/j.tig.

2016.03.004.

53. Hejnol,A., and Martindale, M.Q. (2009) Coordinated

spatial and temporal expression of Hox genes during

embryogenesis in the acoel Convolutriloba longifis-

sura, BMC Biol., 7, 65, doi:10.1186/1741-7007-7-65.

54. Moreno, E., Nadal, M., Baguñà, J., and Martínez, P.

(2009) Tracking the origins of the bilaterian Hox

patterning system: insights from the acoel flatworm

Symsagittifera roscoffensis, Evol. Dev., 11, 574-581,

doi:10.1111/j.1525-142X.2009.00363.x.

55. Moreno,E., De Mulder,K., Salvenmoser,W., Ladurn-

er,P., and Martínez,P. (2010) Inferring the ancestral

function of the posterior Hox gene within the bila-

teria: controlling the maintenance of reproductive

structures, the musculature and the nervous system

in the acoel flatworm Isodiametra pulchra, Evol. Dev.,

12, 258-266, doi:10.1111/j.1525-142X.2010.00411.x.

56. Moreno, E., Permanyer, J., and Martinez, P. (2011)

The origin of patterning systems in bilateria-insights

from the Hox and ParaHox genes in Acoelomorpha,

Genom. Proteom. Bioinform., 9, 65-76, doi: 10.1016/

s1672-0229(11)60010-7.

57. Finnerty, J. R., and Martindale, M.Q. (1999) Ancient

origins of axial patterning genes: Hox genes and

ParaHox genes in the Cnidaria, Evol. Dev., 1, 16-23,

doi:10.1046/j.1525-142x.1999.99010.x.

58. Chiori, R., Jager, M., Denker, E., Wincker, P., Da Sil-

va,C., Le Guyader,H., Manuel,M., and Quéinnec, E.

(2009) Are Hox genes ancestrally involved in axial

patterning? Evidence from the hydrozoan Clytia hemi-

sphaerica (Cnidaria), PLoS One, 4, e4231, doi:10.1371/

journal.pone.0004231.

59. Nong,W., Cao,J., Li,Y., Qu,Z., Sun,J., Swale,T., Yip,

H.Y., Qian, P.Y., Qiu, J.W., Kwan, H.S., Bendena,W.,

Tobe,S., Chan, T.F., Yip, K.Y., Chu, K.H., Ngai, S.M.,

Tsim, K.Y., Holland, P.W.H., and Hui, J.H.L. (2020)

Jellyfish genomes reveal distinct homeobox gene clus-

ters and conservation of small RNA processing, Nat.

Commun., 11, 3051, doi:10.1038/s41467-020-16801-9.

60. Galliot, B., Quiquand, M., Ghila, L., de Rosa, R., Mil-

jkovic-Licina, M., and Chera, S. (2009) Origins of

neurogenesis, a cnidarian view, Dev. Biol., 332, 2-24,

doi:10.1016/j.ydbio.2009.05.563.

61. Faltine-Gonzalez, D., Havrilak, J., and Layden, M. J.

(2023) The brain regulatory program predates cen-

tral nervous system evolution, Sci. Rep., 13, 8626,

doi:10.1038/s41598-023-35721-4.

62. Oren, M., Brickner, I., Appelbaum, L., and Levy, O.

(2014) Fast neurotransmission related genes are ex-

pressed in non nervous endoderm in the sea anem-

one Nematostella vectensis, PLoS One, 9, e93832,

doi:10.1371/journal.pone.0093832.

63. Steinworth, B. M., Martindale, M. Q., and Ryan, J. F.

(2023) Gene Loss may have shaped the Cnidarian and

Bilaterian Hox and ParaHox complement, Genome

Biol. Evol., 15, evac172, doi:10.1093/gbe/evac172.

64. Reggiori, F., and Klionsky, D. J. (2002) Autopha-

gy in the eukaryotic cell, Eukaryot. Cell, 1, 11-21,

doi:10.1128/ec.01.1.11-21.2002.

65. Wada, Y., Sun-Wada, G. H., Kawamura, N., and Aoy-

ama, M. (2014) Role of autophagy in embryogene-

sis, Curr. Opin. Genet. Dev., 27, 60-66, doi: 10.1016/

j.gde.2014.03.010.

66. Tsukamoto,S., Kuma,A., and Mizushima,N. (2008) The

role of autophagy during the oocyte-to-embryo transi-

tion, Autophagy, 4, 1076-1078, doi:10.4161/auto.7065.

67. Banreti,A., Hudry,B., Sass,M., Saurin, A.J., and Gra-

ba, Y. (2014) Hox proteins mediate developmental

and environmental control of autophagy, Dev. Cell,

28, 56-69, doi:10.1016/j.devcel.2013.11.024.

68. Sachs, M., Onodera, C., Blaschke, K., Ebata, K. T.,

Song, J. S., and Ramalho-Santos, M. (2013) Bivalent

chromatin marks developmental regulatory genes in

the mouse embryonic germline in vivo, Cell Rep., 3,

1777-1784, doi:10.1016/j.celrep.2013.04.032.

69. Paul, D., Bridoux, L., Rezsöhazy, R., and Donnay, I.

(2011) HOX genes are expressed in bovine and mouse

oocytes and early embryos, Mol. Reprod. Dev., 78,

436-449, doi:10.1002/mrd.21321.

70. Kondo,M., Yamamoto,T., Takahashi,S., and Taira,M.

(2017) Comprehensive analyses of hox gene expres-

sion in Xenopus laevis embryos and adult tissues, Dev.

Growth Differ., 59, 526-539, doi:10.1111/dgd.12382.

71. Maslakov, G.P., Kulishkin, N. S., Surkova, A.A., and

Kulakova, M. A. (2021) Maternal transcripts of Hox

genes are found in oocytes of Platynereis dumerilii

(Annelida, Nereididae), J.Dev. Biol., 9, 37, doi:10.3390/

jdb9030037.

72. Chipman, A.D., Ferrier, D.E., Brena,C., Qu,J., Hughes,

D.S., Schröder,R., Torres-Oliva,M., Znassi,N., Jiang,H.,

Almeida, F.C., Alonso, C.R., Apostolou,Z., Aqrawi,P.,

Arthur,W., Barna, J.C., Blankenburg, K.P., Brites,D.,

Capella-Gutiérrez,S., Coyle,M., Dearden, P. K., etal.

(2014) The first myriapod genome sequence reveals

conservative arthropod gene content and genome or-

ganisation in the centipede Strigamia maritima, PLoS

Biol., 12, e1002005, doi:10.1371/journal.pbio.1002005.

73. Qu, Z., Nong, W., So, W. L., Barton-Owen, T., Li, Y.,

Leung, T.C.N., Li,C., Baril,T., Wong, A.Y.P., Swale,T.,

Chan, T.F., Hayward,A., Ngai, S.M., and Hui, J.H.L.

(2020) Millipede genomes reveal unique adaptations

during myriapod evolution, PLoS Biol., 18, e3000636,

doi:10.1371/journal.pbio.3000636.

KULAKOVA et al.1000

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

74. Rafiqi, A.M., Rajakumar,A., and Abouheif,E. (2020)

Origin and elaboration of a major evolutionary

transition in individuality, Nature, 585, 239-244,

doi:10.1038/s41586-020-2653-6.

75. Jaramillo, M. L., Ammar, D., Quispe, R. L., Bonat-

to Paese, C. L., Gruendling, A. P., Müller, Y. M., and

Nazari, E.M. (2022) Identification of Hox genes and

their expression profiles during embryonic devel-

opment of the emerging model organism, Macro-

brachium olfersii, J.Exp. Zool. B Mol. Dev. Evol., 338,

292-300, doi:10.1002/jez.b.23118.

76. Ferreira, E.M., Vireque, A.A., Adona, P.R., Meirelles,

F. V., Ferriani, R. A., and Navarro, P. A. (2009) Cyto-

plasmic maturation of bovine oocytes: structural and

biochemical modifications and acquisition of devel-

opmental competence, Theriogenology, 71, 836-848,

doi:10.1016/j.theriogenology.2008.10.023.

77. Sauvegarde, C., Paul, D., Bridoux, L., Jouneau, A.,

Degrelle, S., Hue, I., Rezsohazy, R., and Donnay, I.

(2016) Dynamic pattern of HOXB9 protein localiza-

tion during oocyte maturation and early embryonic

development in mammals, PLoS One, 11, e0165898,

doi:10.1371/journal.pone.0165898.

78. Rinn, J. L., Kertesz, M., Wang, J. K., Squazzo, S. L.,

Xu, X., Brugmann, S. A., Goodnough, L. H., Helms,

J.A., Farnham, P.J., Segal,E., and Chang, H.Y. (2007)

Functional demarcation of active and silent chroma-

tin domains in human HOX loci by noncoding RNAs,

Cell, 129, 1311-1323, doi:10.1016/j.cell.2007.05.022.

79. Iyyappan,R., Aleshkina,D., Zhu,L., Jiang,Z., Kintero-

va, V., and Susor, A. (2021) Oocyte specific lncRNA

variant Rose influences oocyte and embryo develop-

ment, Noncoding RNA Res., 6, 107-113, doi: 10.1016/

j.ncrna.2021.06.001.

80. Kageyama,S., Gunji,W., Nakasato,M., Murakami,Y.,

Nagata,M., and Aoki,F. (2007) Analysis of transcrip-

tion factor expression during oogenesis and preim-

plantation development in mice, Zygote, 15, 117-128,

doi:10.1017/s096719940700411x.

81. Holland, P. W., Booth, H. A., and Bruford, E. A.

(2007) Classification and nomenclature of all hu-

man homeobox genes, BMC Biol., 5, 47, doi:10.1186/

1741-7007-5-47.

82. Rajkovic, A., Pangas, S. A., Ballow, D., Suzumori, N.,

and Matzuk, M. M. (2004) NOBOX deficiency dis-

rupts early folliculogenesis and oocyte-specific gene

expression, Science, 305, 1157-1159, doi: 10.1126/

science.1099755.

83. Jukam, D., Shariati, S. A. M., and Skotheim, J. M.

(2017) Zygotic genome activation in vertebrates,

Dev. Cell, 42, 316-332, doi:10.1016/j.devcel.2017.07.026.

84. Carnesecchi, J., Boumpas, P., van Nierop, Y. S. P.,

Domsch, K., Pinto, H. D., Borges Pinto, P., and

Lohmann, I. (2022) The Hox transcription factor

Ultrabithorax binds RNA and regulates co-transcrip-

tional splicing through an interplay with RNA poly-

meraseII, Nucleic Acids Res., 50, 763-783, doi:10.1093/

nar/gkab1250.

85. Rubin, E., Wu, X., Zhu, T., Cheung, J. C., Chen, H.,

Lorincz, A., Pandita, R. K., Sharma, G. G., Ha, H. C.,

Gasson,J., Hanakahi, L. A., Pandita, T. K., and Suku-

mar, S. (2007) A role for the HOXB7 homeodomain

protein in DNA repair, Cancer Res., 67, 1527-1535,

doi:10.1158/0008-5472.Can-06-4283.

86. Salsi, V., Ferrari, S., Ferraresi, R., Cossarizza, A.,

Grande,A., and Zappavigna,V. (2009) HOXD13 binds

DNA replication origins to promote origin licens-

ing and is inhibited by geminin, Mol. Cell Biol., 29,

5775-5788, doi:10.1128/MCB.00509-09.

87. Primon, M., Hunter, K. D., Pandha, H. S., and Mor-

gan, R. (2019) Kinase regulation of HOX transcrip-

tion factors, Cancers (Basel), 11, 508, doi: 10.3390/

cancers11040508.

88. Lavin, M. F., and Kozlov, S. (2007) ATM activation

and DNA damage response, Cell Cycle, 6, 931-942,

doi:10.4161/cc.6.8.4180.

89. Comelli, L., Marchetti, L., Arosio, D., Riva, S., Abdu-

rashidova,G., Beltram,F., and Falaschi,A. (2009) The

homeotic protein HOXC13 is a member of human

DNA replication complexes, Cell Cycle, 8, 454-459,

doi:10.4161/cc.8.3.7649.

90. Marchetti, L., Comelli, L., D’Innocenzo, B., Puzzi, L.,

Luin, S., Arosio, D., Calvello, M., Mendoza-Maldona-

do,R., Peverali,F., Trovato,F., Riva,S., Biamonti, G.,

Abdurashidova, G., Beltram, F., and Falaschi, A.

(2010) Homeotic proteins participate in the function

of human-DNA replication origins, Nucleic Acids Res.,

38, 8105-8119, doi:10.1093/nar/gkq688.

91. Vieux, K.F., and Clarke, H.J. (2018) CNOT6 regulates a

novel pattern of mRNA deadenylation during oocyte

meiotic maturation, Sci. Rep., 8, 6812, doi: 10.1038/

s41598-018-25187-0.

92. Merabet, S., and Carnesecchi, J. (2024) Hox dosage

and morphological diversification during develop-

ment and evolution, Semin. Cell Dev. Biol., 152-153,

70-75, doi:10.1016/j.semcdb.2022.11.009.

93. Horan, G. S., Ramírez-Solis, R., Featherstone, M. S.,

Wolgemuth, D. J., Bradley, A., and Behringer, R. R.

(1995) Compound mutants for the paralogous

hoxa-4, hoxb-4, and hoxd-4 genes show more com-

plete homeotic transformations and a dose-depen-

dent increase in the number of vertebrae trans-

formed, Genes Dev., 9, 1667-1677, doi: 10.1101/

gad.9.13.1667.

94. Zákány, J., Fromental-Ramain, C., Warot, X., and

Duboule, D. (1997) Regulation of number and size

of digits by posterior Hox genes: a dose-dependent

mechanism with potential evolutionary implica-

tions, Proc. Natl. Acad. Sci. USA, 94, 13695-13700,

doi:10.1073/pnas.94.25.13695.

95. Paul,R., Giraud,G., Domsch,K., Duffraisse,M., Mar-

migère, F., Khan, S., Vanderperre, S., Lohmann, I.,

IRREDUCIBLE COMPLEXITY OF HOX GENE 1001

BIOCHEMISTRY (Moscow) Vol. 89 No. 6 2024

Stoks, R., Shashidhara, L. S., and Merabet, S. (2021)

Hox dosage contributes to flight appendage mor-

phology in Drosophila, Nat. Commun., 12, 2892,

doi:10.1038/s41467-021-23293-8.

96. Joliot, A., Pernelle,C., Deagostini-Bazin, H., and Pro-

chiantz, A. (1991) Antennapedia homeobox peptide

regulates neural morphogenesis, Proc. Natl. Acad. Sci.

USA, 88, 1864-1868, doi:10.1073/pnas.88.5.1864.

97. Lee, E.J., Kim,N., Park, J.W., Kang, K.H., Kim, W.I.,

Sim, N.S., Jeong, C.S., Blackshaw,S., Vidal,M., Huh,

S. O., Kim, D., Lee, J.H., and Kim, J. W. (2019) Glob-

al analysis of intercellular homeodomain protein

transfer, Cell Rep., 28, 712-722.e713, doi: 10.1016/

j.celrep.2019.06.056.

98. Lucas, W. J., Bouché-Pillon, S., Jackson, D. P., Nguy-

en, L., Baker, L., Ding, B., and Hake, S. (1995) Selec-

tive trafficking of KNOTTED1 homeodomain protein

and its mRNA through plasmodesmata, Science, 270,

1980-1983, doi:10.1126/science.270.5244.1980.

99. Prochiantz, A., and Di Nardo, A. A. (2015) Homeo-

protein signaling in the developing and adult ner-

vous system, Neuron, 85, 911-925, doi: 10.1016/

j.neuron.2015.01.019.

100. Spatazza,J., Di Lullo,E., Joliot,A., Dupont,E., Moya,

K.L., and Prochiantz,A. (2013) Homeoprotein signal-

ing in development, health, and disease: a shaking of

dogmas offers challenges and promises from bench

to bed, Pharmacol. Rev., 65, 90-104, doi: 10.1124/

pr.112.006577.

101. Joliot, A.H., Triller,A., Volovitch,M., Pernelle,C., and

Prochiantz, A. (1991) Alpha-2,8-polysialic acid is the

neuronal surface receptor of antennapedia homeo-

box peptide, New Biol., 3, 1121-1134.

102. Merabet,S., Kambris, Z., Capovilla, M., Bérenger, H.,

Pradel, J., and Graba, Y. (2003) The hexapep-

tide and linker regions of the AbdA Hox pro-

tein regulate its activating and repressive func-

tions, Dev. Cell, 4, 761-768, doi: 10.1016/s1534-5807

(03)00126-6.

103. Duffraisse, M., Paul, R., Carnesecchi, J., Hudry, B.,

Banreti,A., Reboulet, J., Ajuria,L., Lohmann, I., and

Merabet, S. (2020) Role of a versatile peptide motif

controlling Hox nuclear export and autophagy in

the Drosophila fat body, J. Cell Sci., 133, jcs241943,

doi:10.1242/jcs.241943.

104. Bardine, N., Lamers, G., Wacker, S., Donow, C.,

Knoechel,W., and Durston,A. (2014) Vertical signal-

ling involves transmission of Hox information from

gastrula mesoderm to neurectoderm, PLoS One, 9,

e115208, doi:10.1371/journal.pone.0115208.

105. Kim, H.T., Kim, S.J., Sohn, Y.I., Paik, S.S., Caplette,R.,

Simonutti,M., Moon, K.H., Lee, E.J., Min, K.W., Kim,

M.J., Lee, D.G., Simeone,A., Lamonerie,T., Furuka-

wa,T., Choi, J.S., Kweon, H.S., Picaud,S., Kim, I.B.,

Shong,M., and Kim, J.W. (2015) Mitochondrial protec-

tion by exogenous Otx2 in mouse retinal neurons, Cell

Rep., 13, 990-1002, doi:10.1016/j.celrep.2015.09.075.

106. Brunet, I., Weinl, C., Piper, M., Trembleau, A., Volo-

vitch, M., Harris, W., Prochiantz, A., and Holt, C.

(2005) The transcription factor Engrailed-2 guides

retinal axons, Nature, 438, 94-98, doi: 10.1038/

nature04110.

107. DuBuc, T.Q., Stephenson, T.B., Rock, A.Q., and Mar-

tindale, M. Q. (2018) Hox and Wnt pattern the pri-

mary body axis of an anthozoan cnidarian before

gastrulation, Nat. Commun., 22, 2007, doi: 10.1038/

s41467-018-04184-x.

108. Schiemann, S.M., Martín-Durán, J.M., Børve,A., Vel-

lutini, B.C., Passamaneck, Y.J., and Hejnol,A. (2017)

Clustered brachiopod Hox genes are not expressed

collinearly and are associated with lophotrochozoan

novelties, Proc. Natl. Acad. Sci. USA, 114, 1913-1922,

doi:10.1073/pnas.1614501114.

109. Salamanca-Díaz, D. A., Calcino, A. D., de Oliveira,

A. L., and Wanninger, A. (2021) Non-collinear Hox

gene expression in bivalves and the evolution of mor-

phological novelties in mollusks, Sci. Rep., 11, 3575,

doi:10.1038/s41598-021-82122-6.

110. Levin,M., Anavy,L., Cole, A.G., Winter,E., Mostov,N.,

Khair, S., Senderovich, N., Kovalev, E., Silver, D. H.,

Feder, M., Fernandez-Valverde, S. L., Nakanishi, N.,

Simmons,D., Simakov, O., Larsson,T., Liu, S. Y., Jer-

afi-Vider, A., Yaniv,K., Ryan, J.F., Martindale, M. Q.,

et al. (2016) The mid-developmental transition and

the evolution of animal body plans, Nature, 531,

637-641, doi:10.1038/nature16994.

Publisher’s Note. Pleiades Publishing remains

neutral with regard to jurisdictional claims in pub-

lished maps and institutional affiliations.