The BTB-ZF transcription factors

Abstract

The BTB-ZF (broad-complex, tramtrack and bric-à-brac - zinc finger) proteins are encoded by at least 49 genes in mouse and man and commonly serve as sequence-specific silencers of gene expression. This review will focus on the known physiological functions of mammalian BTB-ZF proteins, which include essential roles in the development of the immune system. We discuss their function in terminally differentiated lymphocytes and the progenitors that give rise to them, their action in hematopoietic malignancy and roles beyond the immune system.

Introduction

The zinc finger (ZF) motif was first identified in Xenopus laevis transcription factor IIIA protein¹ and was so named because it contained zinc and grasped DNA. The consensus sequence of this particular motif [two cysteines in close proximity to two histidines (C₂H₂)], has come to be recognized as one of the most common DNA-binding motifs, encoded by 3% of all human genes.²

Some of the earliest C₂H₂ ZF genes were found in Drosophila melanogaster, cloned from the developmental mutants tramtrack,³ broad-complex⁴ and bric-à-brac.⁵ The Broad-complex and Tramtrack proteins were noted to share N-terminal homology,⁴ a finding that was later extended to mammalian ZF proteins ZBTB14 and ZBTB25 (respectively known as ZF5 and KUP at the time).⁶ Originally known as the ZiN (Zinc finger N-terminal) motif, it was also recognized to exist in the Vaccinia virus protein KBTB1 (then known as A55).⁶ A longer region of homology was recognized among other Drosophila proteins^5,7 and in a number of poxvirus proteins,⁸ leading to the use of the abbreviations BTB (for Broad-complex, Tramtrack and Bric-à-brac) or POZ (for Poxvirus and Zinc finger) for this newly recognized domain.

The BTB domain is approximately 120 amino acids in length,⁸ typically present in a single copy, and associated with a variety of other domains, including C₂H₂ zinc finger and Kelch domains.⁹ BTB domain-encoding genes are found among eukaryotic and viral genomes yet are absent from prokaryotic genomes (with the exception of an endosymbiotic bacterium¹⁰). The core molecular function of the BTB domain is to promote protein-protein interactions, although the consequences of these interactions will depend upon the broader context of the protein.⁹ The BTB domain of SKP1, for example, recruits target recognition modules to the Cullin E3 ubiquitin ligase complex.^11,12

Of a predicted 156 human genes that encode BTB domain-containing proteins, 49 also possess a series of C₂H₂ zinc fingers (Fig. 1). These proteins are commonly known as BTB-ZF or POK proteins (for POZ/Krüppel-like, after the defining N-terminal POZ domain and the C₂H₂ zinc fingers characteristic of the Drosophila segmentation protein Krüppel).¹³ Apart from appearances in viral genomes, BTB-ZF proteins appear to be restricted to the genomes of higher eukaryotes and are most numerous in vertebrates. One colorful example in Drosophila is the fruitless gene which, when mutated, causes male flies to indiscriminately court both females and males.^14,15

Figure 1. Domain structure and phylogeny of 49 human BTB-ZF proteins. All contain a single BTB domain and between 2 and 14 C₂H₂ ZF domains, with PATZ1 and ZBTB24 also containing an AT-hook domain. Note that ZBTB12B is a pseudogene, ZBTB8OS does not encode a BTB-ZF protein, and that ANKFY1 has a FYVE-type rather than a C₂H₂ ZF domain. Amino acid sequences of each of the BTB-ZF proteins were obtained from Uniprot, and phylogenetic analysis was performed using Phylogeny.fr¹³⁴: (i) Amino acid sequences were aligned with Multiple Sequence Comparison by Log-Expectation (MUSCLE, v. 3.7; default settings). (ii) The maximum likelihood method in PhyML (v. 3.0, aLRT) was used to generate the phylogenetic tree from the multiple sequence alignment. The WAG substitution model was selected assuming an estimated proportion of invariant sites (of 0.000) and 4 gamma-distributed rate categories to account for rate heterogeneity across sites. The gamma shape parameter was estimated directly from the data (gamma = 1.158). Reliability for internal branch was assessed using the Approximate Likelihood Ratio Test (aLRT; SH-Like). (iii) Graphical representation of the phylogenetic tree was performed with TreeDyn (v. 198.3) in the radial style of DrawTree from the PHYLIP package (v. 3.66).

Since the discovery of mammalian BTB-ZF proteins in leukemic translocations, much has been learned about the function of these proteins. For example, the DNA-binding ZF motifs determine sequence specificity, while the BTB domain promotes oligomerization and the recruitment of transcriptional regulators.⁸ These are typically repressors of transcription such as N-CoR (nuclear receptor corepressor) and SMRT (silencing mediator for retinoid and thyroid hormone receptor),^16,17 but also include activators such as p300.¹⁸ Transcriptional regulation by BTB-ZF proteins is often required for normal development of lymphocytes, while other members regulate fertility, skeletal morphogenesis, and/or neurological development (Table 1).

Table 1. BTB-ZF-encoding genes and the phenotypes associated with their mutation.

#	human symbol	mouse symbol	synonyms	ZFs	phenotypes
1	ZBTB1	Zbtb1	ZNF909	8	immune/hematopoietic99
2	ZBTB2	Zbtb2	ZNF437	4
3	ZBTB3	Zbtb3		2
4	ZBTB4	Zbtb4	KAISO-L1, ZNF903	6
5	ZBTB5	Zbtb5		2
6	ZBTB6	Zbtb6	ZID, ZNF482	4
7	ZBTB7A	Zbtb7a	FBI-1, LRF, pokemon, ZBTB7, ZNF857A	4	immune/hematopoietic,92-94 embryogenesis,32 tumorigenesis32
8	ZBTB7B	Zbtb7b	c-Krox, Th-POK, ZBTB15, ZFP67, ZNF857B	4	immune/hematopoietic17,72,73,77,78
9	ZBTB7C	Zbtb7c	ZBTB36, ZNF857C	4	tumorigenesis34
10	ZBTB8A	Zbtb8a	BOZF1, ZBTB8, ZNF916A	2
11	ZBTB8B	Zbtb8b	ZNF916B	2
12	ZBTB9	Zbtb9	ZNF919	2
13	ZBTB10	Zbtb10	RINZF	2
14	ZBTB11	Zbtb11	ZNF913	12
15	ZBTB12	Zbtb12	G10, NG35, Bat9	4
16	ZFP161	Zfp161	ZBTB14, ZNF478, ZF5	5
17	ZBTB16	Zbtb16	PLZF, ZNF145	9	immune/hematopoietic,79-84,91 tumorigenesis,21,23 skeleton,103,104,106 reproductive103,107,108
18	ZBTB17	Zbtb17	MIZ1, pHZ-67, ZNF151, ZNF60	13	immune/hematopoietic,96,97 embryogenesis,95 integument136
19	ZNF238	Zfp238	C2H2–171, RP58, TAZ-1, ZBTB18	4	nervous120
20	PATZ1	Patz1	MAZR, PATZ, RIAZ, ZBTB19, ZNF278, ZSG	7	immune/hematopoietic,76 reproductive112
21	ZBTB20	Zbtb20	DPZF, ODA-8S, ZNF288	5	nervous,121 liver/biliary122,123
22	ZNF295	Zfp295	ZBTB21	8
23	ZBTB22	Zbtb22	BING1, fruitless, ZBTB22A, ZNF297, ZNF297A	3
24	GZF1	Gzf1	ZBTB23, ZNF336	10
25	ZBTB24	Zbtb24	BIF1, PATZ2, ZNF450	8	immune/hematopoietic127,128
26	ZBTB25	Zbtb25	KUP, ZNF46	2
27	ZBTB26	Zbtb26	ZNF481	4
28	BCL6	Bcl6	BCL5, BCL6A, LAZ3, ZBTB27, ZNF51	6	immune/hematopoietic,43-45,49-58,137 tumorigenesis,24-26,28 reproductive,113 skeleton64
29	BCL6B	Bcl6b	BAZF, ZBTB28, ZNF62	5	immune/hematopoietic89,90,137
30	HIC1	Hic1	ZBTB29, ZNF901	5	embryogenesis,124 tumorigenesis37
31	HIC2	Hic2	HRG22, ZBTB30, ZNF907	5
32	MYNN	Mynn	SBBIZ1, ZBTB31, ZNF902	8
33	ZBTB32	Zbtb32	PLZP, FAXF, FAZF, Rog, TZFP, ZNF538	3	immune/hematopoietic87,88
34	ZBTB33	Zbtb33	KAISO, ZNF-kaiso, ZNF348	3	tumorigenesis33
35	ZBTB34	Zbtb34	ZNF918	3
36	ZBTB37	Zbtb37	ZNF908	3
37	ZBTB38	Zbtb38	CIBZ, ZNF921	10
38	ZBTB39	Zbtb39	ZNF922	8
39	ZBTB40	Zbtb40	ZNF923	12
40	ZBTB41	Zbtb41	FRBZ1, ZNF924	14
41	ZBTB42	Zbtb42	ZNF925	4
42	ZBTB43	Zbtb43	ZBTB22B, ZNF-X, ZNF297B	3
43	ZBTB44	Zbtb44	BTBD15, ZNF851	4
44	ZBTB45	Zbtb45	ZNF499	4
45	ZBTB46	Zbtb46	BTBD4, RINZF, ZNF340, zDC	2	immune/hematopoietic100
46	ZBTB47	Zfp651	ZNF651	9
47	ZBTB48	Zbtb48	HKR3, ZNF855	11
48	ZBTB49	Zbtb49	ZNF509	7
49	ZNF131	Zfp131		6

Hematopoietic Malignancy

The first mammalian BTB-ZF proteins to be characterized in detail emerged from the study of chromosomal translocations in human leukemias. Of the translocations associated with acute promyelocytic leukemia, all of them involve the retinoic acid receptor α gene (RARA) on chromosome 17. In 95% of cases, RARA is fused to the promyelocytic leukemia gene on chromosome 15; in a subset of the remaining cases, RARA forms a fusion with the promyelocytic leukemia zinc finger gene (PLZF, now known as ZBTB16) on chromosome 11.¹⁹ The t(11;17) translocation has the worst prognosis of all, since it is the only one resistant to the standard treatment for acute promyelocytic leukemia: all-trans retinoic acid therapy.²⁰

ZBTB16 encodes a number of ZF motifs,²¹ and, unlike many other translocations, both ZBTB16 reciprocal translocation products are required for oncogenesis. RARα-PLZF by itself is not oncogenic in mice, and although PLZF-RARα causes a chronic myelogenous leukemia-like disease in isolation,²² RARα-PLZF is required for the development of a disease more similar to acute promyelocytic leukemia.²³ The action of RARα-PLZF is likely to be a dominantly interfering one, since deletion of ZBTB16 has the same effect in the context of PLZF-RARα.²³

B cell lymphoma is a second example of a BTB-ZF gene rearrangement associated with hematopoietic malignancy. ~40% of diffuse large B cell lymphomas, as well as 5–10% of follicular lymphomas, are associated with translocations of chromosome 3q27 within the region of an actively transcribed gene. This gene, named B cell lymphoma 6 (BCL6) after the disease associated with it, encodes a number of ZF domains;^24-26 only later was it realized that both BCL6 and PLZF also shared a BTB domain.^7,8 The majority of BCL6 breakpoints occur around its noncoding first exon, leading to promoter substitution and deregulated expression of an otherwise normal BCL6 protein,²⁷ which is sufficient to cause B cell lymphoma when expressed in mice.²⁸ Following tyrosine kinase inhibitor treatment, BCL6 is also overexpressed in BCR-ABL1 acute lymphoblastic leukemia cells at levels similar to those seen in B cell lymphomas with BCL6 translocations.²⁹ BCL6 is therefore an attractive target for the treatment of both B cell lymphoma and BCR-ABL1 leukemia. Targeting the BTB domain of BCL6 by peptide inhibition may be one approach, and has been shown to induce cell cycle arrest and apoptosis by disrupting corepressor recruitment.³⁰ BTB-specific interference might be equally useful for the treatment of other BTB-ZF-associated malignancies.

The BTB-ZF leukemia-related factor (LRF) protein, encoded by ZBTB7A (a paralog of ZBTB7B and ZBTB7C), is known to interact with BCL6³¹ and can induce T cell leukemia in mice when overexpressed by an lck/Eμ enhancer/promoter.³² LRF-deficient mouse fibroblasts are also resistant to oncogene-mediated transformation.³² Similarly, deletion of Zbtb33 can delay tumorigenesis,³³ while the ZBTB7C protein is required for fibroblast proliferation and promotes tumorigenesis when overexpressed.³⁴

HIC1 (hypermethylated in cancer 1) is a candidate tumor suppressor gene underexpressed in many cancers,³⁵ including hematological ones.³⁶ Heterozygous Hic1-knockout mice develop a variety of neoplasms with age (lymphomas in ~25% of cases)³⁷ and exhibit accelerated tumorigenesis on a variety of sensitized backgrounds,^38-40 presumably through SIRT1-mediated inactivation of p53.⁴¹ Finally, while not a tumor suppressor itself, FAZF is an interacting partner of both Fanconi anemia group C protein (FANCC) and PLZF⁴² and may, therefore, play a role in tumorigenesis of Fanconi anemia and/or acute promyelocytic leukemia.

Lymphopoiesis and Immunity

BCL6: A master regulator of the germinal center response.

As one of the first mammalian BTB-ZF genes to be discovered, Bcl6 was also one of the first to be targeted in the mouse genome. Germinal centers were conspicuously absent from Bcl6 mutant mice due in part to a cell-intrinsic failure of germinal center B cell differentiation.^43-45 Immunoglobulin class switching and somatic hypermutation, which are frequent events in germinal center B cells, would otherwise induce a DNA damage-associated apoptotic response. BCL6 allows germinal center B cells to tolerate these events by actively repressing transcription of TP53, the DNA damage sensor ATR and the cell cycle arrest gene CDKN1A.^46-48 Given that p53 is so frequently inactivated in cancer, these findings also offered an explanation for how BCL6 activation could lead to B cell lymphoma.

More recently it has become clear that BCL6 is also required for the differentiation of follicular helper T cells, which support germinal center formation,^49-51 and for follicular regulatory T cells and follicular helper NKT cells.^52-54 Several other hematopoietic lineages require BCL6 for their development, including memory T cells (CD4⁺ and CD8⁺)^55-57 and conventional dendritic cells (CD4⁺ and CD8α⁺).⁵⁸

BCL6 also actively counteracts the differentiation of other cell lineages. While BCL6 is necessary for the formation of germinal center B cells, their progression to immunoglobulin-secreting plasma cells requires the transcriptional regulator BLIMP-1 (encoded by Prdm1).^59,60 Prdm1 itself is directly repressed by BCL6;^61-63 only after Bcl6 expression is extinguished can plasma cell differentiation occur. Osteoclasts, which are derived from hematopoietic precursors in the bone marrow, also depend upon the Bcl6/Prdm1 axis. BCL6 is a suppressor of osteoclastogenesis by default, while BLIMP-1 promotes it, likely via direct repression of the Bcl6 promoter by BLIMP-1.⁶⁴

Another defining phenotype of the original Bcl6-knockout mice was their development of a lethal Th2-type inflammatory disease.^43,44 Neither IL-4 nor STAT6 was required for disease development,⁶⁵ which was driven by non-lymphoid cells,⁶⁶ although BCL6 was required for STAT6-dependent IgE class switching.⁶⁷ The critical inflammatory cell type appears to be macrophages, in which BCL6-mediated repression of proinflammatory chemokines and cytokines is lost,^66,68 driving macrophage proliferation⁶⁸ and promoting Th17 differentiation.⁶⁹

Th-POK: Thymic lineage commitment and effector T cell function.

One of the key stages of T cell lineage commitment occurs in the thymus, where CD4⁺CD8⁺ thymocytes diverge into either the CD4⁺ helper or CD8⁺ cytotoxic lineage. The molecular basis of this choice had been a mystery for many years, until the serendipitous discovery of a spontaneous mouse mutant, hd (helper T cell deficient). hd mice entirely lack CD4 single positive cells in the thymus⁷⁰ and T cells which would otherwise be destined for the CD4 lineage (e.g., those with an MHC class II-restricted TCR transgene or those developing in MHC class I-deficient hosts) are diverted to the CD8 lineage in hd mice, implying that CD8 was the default lineage in the absence of the hd gene product.⁷¹ The hd mutation was eventually revealed to be a missense allele of Zbtb7b, encoding the BTB-ZF protein Th-POK⁷² (also known as cKrox).⁷³ Zbtb7b, which is normally silenced by Runx transcription factor complexes in CD8 cells,⁷⁴ does not appear to be essential for CD4 specification but is required to prevent Runx-dependent commitment to the CD8 lineage.⁷⁵ PATZ1 (POZ-, AT hook- and zinc finger-containing protein 1), another BTB-ZF protein, also contributes to silencing of the Zbtb7b locus in CD8⁺ cells.⁷⁶

Th-POK is required for the normal maturation of invariant NKT cells^17,77 as well as for the optimal development of γδ T cells.⁷⁸ PLZF is also essential for the development of a variety of innate T cells, including effector NKT cells^79-81 and innate γδ T cells.^82-84 The effect of PLZF on innate γδ T cell development is not an intrinsic one, however, and instead requires the secretion of IL-4 by PLZF⁺ NKT cells.^85,86

At least two BTB-ZF proteins have been shown to be involved in TCR-mediated proliferation, with varying roles. PLZP-deficient T cells proliferate more in response to TCR and cytokine receptor stimulation,^87,88 while anti-CD3-induced proliferation of naïve, but not memory CD4⁺ T cells, is partially reduced in BCL6B-deficient mice.⁸⁹ Conversely, the recall response of virus-specific memory CD8⁺ T cells is BCL6B-dependent, while their primary response is not.⁹⁰ In the broader context of viral infection, PLZF plays an important role in the induction of type I interferon-inducible gene expression. Through an interaction with the histone deacetylase HDAC1 and the promyelocytic leukemia protein PML, PLZF binds to the promoter regions of interferon-inducible genes and promotes their expression.⁹¹ Accordingly, PLZF-deficient NK cells have impaired function, and PLZF-deficient mice show a heightened sensitivity to viral infection.⁹¹

LRF and MIZ1: T vs B lineage commitment and beyond.

T and B lymphocytes arise from common lymphoid precursors in the bone marrow. Within these cells, the action of at least two BTB-ZF proteins influences the choice of lineage: LRF, encoded by Zbtb7a, and MIZ1, encoded by Zbtb17. While germline deletion of Zbtb7a is lethal to the developing embryo,^32,92 fetal liver chimera and conditional deletion experiments established that LRF was essential for fetal and adult B cell development.⁹² In the conditional LRF mutants, failure of B cell development was associated with ectopic development of CD4⁺CD8⁺ cells in the bone marrow. Treating Zbtb7a mutant progenitors with a Notch1 inhibitor could partially correct this developmental imbalance, implying that LRF promotes B cell lineage commitment by counteracting Notch1.⁹²

LRF is important beyond the T/B lineage split, with B cell-specific deletion of Zbtb7a resulting in the generation of more marginal zone B cells but fewer follicular and germinal center B cells.⁹³ The embryonic lethality of LRF-deficient mice is also associated with profound anemia.^32,92 GATA1, a key transcriptional regulator of erythropoiesis, was found to activate the expression of Zbtb7a, which, in turn, suppressed expression of the gene encoding the proapoptotic protein Bim.⁹⁴ Accordingly, deficiency of Bim could correct the anemia in Zbtb7a mutant embryos and prolong their survival in utero.⁹⁴

Germline deletion of Zbtb17 (encoding the MIZ1 protein) is also lethal in utero, but at a much earlier stage than Zbtb7a (E7.5) due to a failure of gastrulation.⁹⁵ Zbtb17 mutant cells can nevertheless contribute to myeloid and erythroid lineages in ES cell chimeras yet were not represented in T or B lymphocytes.⁹⁶ Hematopoietic conditional deletion of Zbtb17 using a Vav-cre transgene also led to deficiencies of T and B cells, with respective blocks at the early T cell progenitor and pre-pro-B stages of development due to insufficient IL-7R signaling.^96,97 The block in IL-7R signaling could be explained in part by the overexpression of Socs1, which MIZ1 would otherwise repress. Inhibition of Socs1 could rescue lymphoid development to a small extent in vitro, with more substantial improvements seen after Bcl2 transgene expression (in combination with Ebf1 overexpression for B cells).^96,97

ZBTB1: A determinant of lymphoid development.

One of the more recently characterized BTB-ZF members is ZBTB1. Like many other BTB-ZF proteins, ZBTB1 is a potent transcriptional repressor,⁹⁸ whose localization and repressive activity is regulated by SUMOylation.⁹⁸ Forward genetic screening in mice revealed that a missense mutation in the BTB domain of ZBTB1 (C74R) leads to a profound T cell deficiency, with milder deficiencies of B and NK cells.⁹⁹ When placed in competition with wild-type cells, however, Zbtb1 mutant cells fail to generate any lymphocytes at all, while myelopoiesis is unperturbed.⁹⁹

ZBTB46: A marker of conventional DCs.

One of the most highly expressed genes in mouse conventional dendritic cells (cDCs) is Zbtb46, a BTB-ZF gene also expressed in endothelial cells and erythroid precursors.¹⁰⁰ Knock-in alleles at the Zbtb46 locus have allowed the creation of cDC-specific reporters, distinguishing them from other CD11c⁺ phagocytes and allowing lineage-specific ablation.^100,101 Zbtb46 is not essential for their differentiation, but may be important for their full functional maturation, as expression of the receptors for G-CSF and leukemia inhibitory factor persist in its absence.¹⁰⁰ Retroviral overexpression of Zbtb46 promotes cDC development at the expense of granulocytes and can even overcome the cDC developmental block in Irf8 mutant progenitors.¹⁰⁰ This effect was dependent upon the tandem ZF motifs, as a ZF deletion mutant could not enhance cDC development.

Functions of BTB-ZF Proteins Beyond Hematopoiesis

Limb patterning.

Long before BTB-ZF proteins were known, the effects of a particular BTB-ZF mutation had been closely studied. Green’s luxoid is a spontaneous mouse mutant characterized by skeletal abnormalities similar to the luxate mutant.^102,103 The observed limb patterning defects in this model are a consequence of a nonsense mutation in Zbtb16 (encoding PLZF), are recapitulated in a Zbtb16-knockout strain¹⁰⁴ and are influenced by the transcriptional regulator Gli3.¹⁰⁵ Similar skeletal abnormalities, including an absence of thumbs, were observed in a PLZF-deficient patient, who inherited an 11q23 deletion in trans with a recessive missense mutation of ZBTB16.¹⁰⁶

Fertility.

A second defining feature of luxoid mutants was their infertility, which has since been explained by a failure of spermatogonial stem cell self-renewal,^107,108 potentially due to the inhibition of mTORC1¹⁰⁹ or the stem cell antigen Kit.¹¹⁰ These pleiotropic effects of PLZF are likely to be due to BTB-mediated interactions with distinct proteins, as they can be separated by a point mutation in the BTB domain.¹¹¹ Genital hypoplasia is also a feature of human PLZF deficiency,¹⁰⁶ suggesting the conservation of both functions in man. Similar to PLZF, loss of PATZ1 causes male infertility due to a loss of spermatogonial stem cells;¹¹² germ cell apoptosis is also a feature of BCL6 mutant males.¹¹³ Sterility is also a feature of at least two Drosophila BTB-ZF mutants, ken and barbie, characterized by loss of terminalia^114,115 and mutations in the fly ortholog of ZBTB22, fruitless, which causes male sterility due to behavioral defects, rather than a failure to develop germ cells or genitalia.^14,15

Neurological and other developmental roles.

A more common function for Drosophila BTB-ZF proteins is the regulation of neuronal development. Tramtrack itself is required to prevent transformation of neuronal support cells into neurons,¹¹⁶ while fruitless, abrupt, lola and chinmo are all examples of Drosophila BTB-ZF mutants that affect neurological development when mutated.^14,117-119 Similarly, mutant alleles of mouse Zfp238 exhibit cortical and hippocampal dysplasia.¹²⁰ Zbtb20 is required for the development of clusters of hippocampal neurons¹²¹ as well as for glucose metabolism and development of the liver,¹²² where it suppresses transcription of the hepatocyte proliferation gene AFP.¹²³ Miller-Dieker syndrome, caused by a multigene chromosomal deletion spanning HIC1, is associated with mental retardation and a variety of developmental abnormalities, some of which are recapitulated in homozygous Hic1 mutant embryos.¹²⁴ KAISO, encoded by Zbtb33, is redundant during embryonic development in mice,³³ yet it is essential in Xenopus.¹²⁵ This may be explained in part by the compensatory activity of ZBTB4 and ZBTB38, which, like KAISO, bind to methylated DNA and repress transcription.¹²⁶

BTB-ZF proteins of obscure physiological function.

The vast majority of BTB-ZF genes have not been studied in germline mutant mice, yet their expression patterns hint that many others may be important for hematopoiesis and other developmental pathways (Fig. 4). One alluring candidate is ZBTB24, which is mutated in immunodeficiency, centromeric instability and facial anomalies syndrome type 2 (ICF2).^127,128 ZBTB24 mutant patients often succumb to recurrent respiratory or gastrointestinal infections, although the nature of their immunodeficiency has only been linked to a deficiency of immunoglobulins, usually in the presence of normal T and B lymphocyte frequencies.¹²⁹ ICF2 patients share features of their disorder with ICF1 patients, who carry mutations in the DNA methyltransferase 3B gene (DNMT3B).¹³⁰ One characteristic feature is centromeric instability, commonly associated with hypomethylated juxtacentromeric regions of chromosomes 1, 9 and 16. How this leads to the clinical features of ICF is unknown, as is the means by which ZBTB24 prevents chromosomal instability. Along with PATZ1, ZBTB24 is also one of only two mammalian BTB-ZF proteins recognized to contain an AT-hook domain, a DNA-binding motif capable of interacting with the minor groove of AT-rich sequences.¹³¹

Figure 4. Collective display of mouse BTB-ZF gene expression data. Relative tissue-specific transcript expression data were obtained from BioGPS¹³⁵ and are representative of multiple transcript probes where appropriate. Gray columns indicate hematopoietic tissues, while black columns are non-hematopoietic. pDC, plasmacytoid dendritic cell; BMDM, bone marrow-derived macrophage; LPS, lipopolysaccharide; GMP, granulocyte-monocyte progenitor; MEP, megakaryocyte-erythrocyte progenitor; HSC, hematopoietic stem cell.

Given that mutant alleles of almost every BTB-ZF gene already exist in mouse ES cells¹³² and in chemically-induced mutant archives,¹³³ it is only a matter of time before we learn the non-redundant functions of every BTB-ZF protein. Rare loss-of-function alleles of the same genes will also exist in the human population and may emerge in homozygous or compound heterozygous form as genomes of human patients are sequenced at a rapidly increasing pace.

Summary

A host of physiological functions have emerged for the mammalian BTB-ZF proteins, from embryogenesis to the development of specialized lymphocyte effector cells. (Fig. 2) As new terminal branches of lymphocyte differentiation are revealed, so too are new functions for BTB-ZF proteins. Transcriptional suppression appears to be a unifying functional trait, (Fig. 3) orchestrated by BTB-mediated recruitment of chromatin remodeling factors to ZF-defined recognition sites, and there is increasing evidence that individual BTB-ZF proteins can influence multiple developmental pathways. The functions of most remain unknown, and for the ones that are known, the proteins they interact with and the genes they regulate are mostly a mystery.

Figure 2. Schematic of BTB-ZF protein functions in hematopoiesis and immunity. Labels outside cells represent a requirement for the BTB-ZF protein during development. Labels inside cells represent a requirement in effector function. See text for additional details.

Figure 3. Transcriptional circuits involving mammalian BTB-ZF proteins. Examples shown are from pathways established in mammalian primary cells, question marks indicate that a direct interaction is yet to be demonstrated. See text for additional details.