A comprehensive catalog of human KRAB-associated zinc finger

A comprehensive catalog of human KRAB-associated

zinc finger genes: Insights into the evolutionary

history of a large family of transcriptional repressors

Stuart Huntley,

Daniel M. Baggott,

Aaron T. Hamilton,

Mary Tran-Gyamfi,

Shan Yang,

Joomyeong Kim,

Laurie Gordon,

Elbert Branscomb,

and Lisa Stubbs

1,4

Genome Biology and

Microbial Systems Divisions, Biosciences, Lawrence Livermore National Laboratory,

Livermore, California 94550, USA

Krüppel-type zinc finger (ZNF) motifs are prevalent components of transcription factor proteins in all eukaryotes.

KRAB-ZNF proteins, in which a potent repressor domain is attached to a tandem array of DNA-binding zinc-finger

motifs, are specific to tetrapod vertebrates and represent the largest class of ZNF proteins in mammals. To define the

full repertoire of human KRAB-ZNF proteins, we searched the genome sequence for key motifs and then constructed

and manually curated gene models incorporating those sequences. The resulting gene catalog contains 423

KRAB-ZNF protein-coding loci, yielding alternative transcripts that altogether predict at least 742 structurally

distinct proteins. Active rounds of segmental duplication, involving single genes or larger regions and including both

tandem and distributed duplication events, have driven the expansion of this mammalian gene family. Comparisons

between the human genes and ZNF loci mined from the draft mouse, dog, and chimpanzee genomes not only

identified 103 KRAB-ZNF genes that are conserved in mammals but also highlighted a substantial level of

lineage-specific change; at least 136 KRAB-ZNF coding genes are primate specific, including many recent duplicates.

KRAB-ZNF genes are widely expressed and clustered genes are typically not coregulated, indicating that paralogs

have evolved to fill roles in many different biological processes. To facilitate further study, we have developed a

Web-based public resource with access to gene models, sequences, and other data, including visualization tools to

provide genomic context and interaction with other public data sets.

[Supplemental material is available online at www.genome.org.]

The human genome contains ∼30,000 genes (Lander et al. 2001;

Venter et al. 2001) including at least 2000 loci encoding tran-

scription factor proteins (TFs) (Messina et al. 2004). The C2H2, or

Krüppel-type zinc finger (ZNF), is the most common DNA-

binding motif found in eukaryotic TF proteins; ZNF proteins typi-

cally contain multiple C2H2 motifs joined together in tandem

arrays. Proteins of this type are ancient and numerous, encoded

by large and diverse gene families in all eukaryotic genomes.

Subfamilies of distinct structure and function have arisen in dif-

ferent evolutionary lineages defined by the types of chromatin

interaction modules, or effector motifs, included in the protein

(Knochel et al. 1989; Bellefroid et al. 1991; Chung et al. 2002; for

reviews, see Collins et al. 2001; Huntley et al., in press).

At least one-third of mammalian ZNF proteins include an

effector motif called the Krüppel-associated box, or KRAB, which

serves to recruit histone deacetylase complexes to regions sur-

rounding the DNA-binding sites (Bellefroid et al. 1991; Friedman

et al. 1996; Pengue and Lania 1996; Abrink et al. 2001; Ayya-

nathan et al. 2003). KRAB-associated ZNF (KRAB-ZNF) proteins

thus function as potent transcriptional repressors (Margolin et al.

1994; Friedman et al. 1996; Ayyanathan et al. 2003). Proteins

combining KRAB and ZNF domains are specific to tetrapod ver-

tebrates, but the family has expanded dramatically to include

hundreds of members in mammals (Bellefroid et al. 1995; Loo-

man et al. 2002; Hamilton et al. 2003). Many KRAB-ZNF

loci reside in familial gene clusters, indicating that the family

has evolved primarily through tandem in situ duplication

(Bellefroid et al. 1995; Collins et al. 2001; Huntley et al., in press).

Recent studies have shown that paralogs diversify through

structural changes in the zinc finger arrays that are driven

by positive selection (Hamilton et al. 2003, 2006; Shannon

et al. 2003; Schmidt and Durrett 2004); since even subtle alter-

ations in zinc-finger array structure can yield proteins with dis-

tinct DNA recognition specificities (Miller and Pabo 2001; Krebs

et al. 2005), this pattern of divergence suggests an active selection

for novel transcription factors with altered regulatory properties.

However, since most KRAB-ZNF genes remain completely un-

characterized, it has not been possible to examine the evolution-

ary histories or functional diversity of this large gene family in

depth.

To derive a complete catalog of the KRAB-ZNF gene family,

we computationally analyzed and manually curated all segments

of the human genome containing KRAB and ZNF domains. These

efforts revealed 423 protein-coding genes with alternative tran-

scripts that predict the existence of at least 742 distinct proteins,

as well as 341 pseudogene sequences. Analyses of this gene set

and comparisons to predicted genes in other mammalian ge-

nomes permitted a genome-wide assessment of the mechanisms

through which this gene family has evolved. Results of this study

can be accessed from a public Web site (http://znf.llnl.gov/) that

will be updated as additional data becomes available.

Present address: Louisiana State University, Baton Rouge, LA.

Corresponding author.

E-mail [email protected]; fax (925) 422-2099.

Article published online before print. Article and publication date are at http://

www.genome.org/cgi/doi/10.1101/gr.4842106.

Resource

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

Results

Assembling the human KRAB-ZNF catalog and Web-based

resource

KRAB-ZNF genes exist as simple modular structures with one or

more KRAB-effector domains and a tandem array of zinc-finger

motifs encoded within distinct 5⬘ and 3⬘ exons, respectively

(Shannon and Stubbs 1998). In addition to the functionally

dominant KRAB-A motif, many genes also encode modulating

motifs such as KRAB-B (Bellefroid et al. 1991), a novel KRAB-B

variant we will refer to as KRAB-BL (KRAB-BL exons are 30 bp

larger than KRAB-B exons, extending in the 3⬘ direction), KRAB-b

(Mark et al. 1999), or KRAB-C (Looman et al. 2004). A small

number of KRAB-ZNF genes also encode a second vertebrate-

specific effector called SCAN (Sander et al. 2003). To catalog the

complete gene family, we employed profile Hidden Markov

Model (HMM) software (HMMER 2.3, http://hmmer.wustl.edu/)

to generate profile HMMs for KRAB, SCAN, and Krüppel-type finger

motifs and to scan the human genome for matching sequences.

Based on RNA evidence and HMM-identified motifs, we gen-

erated models for alternate transcripts arising from each locus

and identified overlaps with publicly available known and pre-

dicted genes. Three hundred thirty-four HMM-based models

overlapped known loci; manual annotation produced 669 tran-

script models for these genes, including 495 models that we ex-

tended or corrected and 157 public models that were not modi-

fied. In addition to the known genes, we identified 89 KRAB-ZNF

loci capable of encoding full-length proteins that are not de-

scribed in public databases (see Methods). Altogether we anno-

tated 423 loci encoding proteins with both effector (KRAB and/or

SCAN) and zinc-finger domains (Table 1; Supplemental Table S1),

for simplicity we will hereafter refer to the collection as KRAB-

ZNF genes.

In addition to KRAB-ZNF loci, we also identified 254 genes

with noncanonical structures, e.g., encoding ZNF-only, KRAB-

only, or SCAN-only proteins as their only potential protein prod-

uct. Loci of this type were annotated as genes only when sup-

ported by mRNA evidence; these genes may indeed correspond to

functional family members. However, in the following discus-

sions we will focus primarily on the 423 loci capable of encoding

proteins with both effector (SCAN, KRAB, or both) and zinc-

finger domains. To publicly share the data arising from this

analysis, we created a Web-based resource (http://znf.llnl.gov/)

that provides access to full descriptions and sequences of all cu-

rated KRAB-ZNF gene models and pseudogene loci. Interfaces for

searching and browsing the database, including an added track

within the UCSC Genome Browser (Kent et al. 2002), are also

provided, along with a set of downloadable data files describing

the models, motifs identified, and HMMER matrices. We will

continue to expand and update the Web resource as new data

regarding these loci are made available.

Alternate splicing and pseudogenes

Based on gene models, we identified 818 different transcripts

encoding 742 structurally distinct proteins from the 423 KRAB-

ZNF genes. The most prevalent class of transcripts includes KRAB-

A, KRAB-B, and ZNF-encoding exons, but proteins with many

different structures were also predicted. For example, alternate

transcripts encoding KRAB-only or fingers-only proteins, or pro-

teins that include only one of several encoded effectors, are fre-

quently generated from the KRAB-ZNF gene sets, as has been

described previously for specific genes (Table 1; Supplemental

Table S1; Bellefroid et al. 1993; Wu et al. 2003; Oh et al. 2005).

Predicted KRAB-ZNF proteins with 2–40 tandem zinc-fingers are

encoded by the collection of human genes, with a median num-

ber of 12 ZNF motifs per gene. SCAN-containing proteins typi-

cally include a smaller number of zinc finger motifs than other

proteins in this family (Table 1). This observation suggests that

the SCAN proteins might typically recognize shorter DNA-

binding motifs, an interpretation that is interesting in light of

the fact that many SCAN-ZNF proteins bind DNA as homodimers

(Sander and Morris 2002).

We also identified 227 gene fragments and 39 full-length

pseudogenes, based on evidence of multiple stop codons, frame-

shifts, and lack of proper splice junctions. Sequence comparisons

confirmed that most pseudogenes arose from neighboring loci by

partial-gene duplication events (data not shown), although gene

remnants may also be left behind after lineage-specific deletions

(Hamilton et al. 2003; Shannon et al. 2003). We also found evi-

dence of 75 processed KRAB-ZNF pseudogene sequences. Three of

these processed pseudogenes maintain open reading frames po-

tentially capable of encoding functional proteins: LLNL1071

(HSA3, 32Mb), LLNL1040 (HSA9, 35Mb), and LLNL973 (HSA12,

132Mb). All other processed loci correspond to degraded, non-

functional copies.

Gene clustering and evolution

We counted 65 KRAB-ZNF, SCAN-ZNF, and mixed gene clusters

in the human genome (for cluster definition and criteria, see

Methods) (Supplemental Table 1). A total of 384 KRAB-ZNF genes

reside in these clusters; the remaining 39 loci were classified as

isolated singleton genes (Table 1). Most of the genes are concen-

trated in 25 major clusters, the largest of which are located on

HSA19 (Table 2). A few KRAB-ZNF gene clusters are well con-

Table 1. Characteristics of KRAB-ZNF gene family members

Locus type

No.

Found Clustered Singletons

Effector-only

transcripts

ZNF-only

transcripts

Median no.

fingers

Conserved in

chimpanzee

Conserved

in mouse

Conserved

in dog

KRAB-A 99 89 10 9 27 12 94 23 42

KRAB-A-B 214 193 21 38 39 12 198 51 120

KRAB-A-BL 11 11 0 2 3 15 10 1 0

KRAB-A-b 11 11 0 1 3 15 11 0 4

KRAB-A-C 31 31 0 1 3 14 29 1 3

SCAN-KRAB-A

25 21 4 6 0 8 25 19 22

SCAN 32 28 4 9 4 5.5 32 17 26

Total 423 384 39 66 79 12 399 112 217

Includes all SCAN-KRAB-A genes with and without additional modulating motifs.

Huntley et al.

670 Genome Research

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

Table 2. Properties of major human KRAB-ZNF clusters and related regions in other mammalian genomes

Human Chimpanzee Dog Mouse

Gene types

Cluster

location Coordinates

Genes

All loci

Coordinates Orthologs All loci Coordinates Orthologs All loci Coordinates Orthologs All loci

SA/AB/AC 1q44 chr1:243 6 6 chr1:227–228 6 6 chr8:4 1 1 chr11:59 2 2

AB 3p22.1a chr3:40 3 4 chr2:41 3 4 chr23:12–13 3 3 — 0 0

SA/A 3p21.32-p21.31 chr3:44 4 12 chr2:45–46 3 10 chr23:4–5 2 3 chr9:123 2 6

chr2_random 1 1 chrUn:14 1 1

A/AB 4p16.3 chr4:0–0.5 4 7 chr3:0.5 4 8 — 0 0 — 0 0

A/AB 5q35.3 chr5:178 6 6 chr4:185 6 6 chr11:9 6 6 chr11:50 6 6

S/SA 6p22.1b chr6:28 9 21 chr5:28–29 9 17 chr35:28 4 12 chr13:20–21 4 7

A/AB 7q11.21a chr7:62 9 17 chr6:64–65 7 18 — 0 0 — 0 0

SA/AB 7q22.1a chr7:98 5 6 chr6:100. 4 6 chr6:12 4 7 chr5:144 5 5

chrUn:130 1 1

A/AB 7q36.1a chr7:148–149 7 13 chr6:151 7 17 chr16:16 7 12 chr6:48 5 8

AB 8q24.3c chr8:145–146 6 12 chr7:149 5 14 chr13:41 4 6 chr15:76 3 6

chrUn:60 1 1

AB 10p11.21 chr10:38 4 6 chr8_random 2 2 chr4:3 2 5 chr6:118 2 3

chr8:42–43 2 4

AB 12q24.33 chr12:132 7 7 chr10:134–135 6 7 chr26:3 6 8 chr5:109 1 9

chr10_random 1 1

SA/AB 16p13.3 chr16:3 9 9 chr18:3 9 9 chr6:39 7 9 chr17:21 3 3

chr16:3 3 3

AB 16p11.2a chr16:30 5 9 chr18:31 4 7 chr6:18 4 10 chr7:121 3 12

chr18_random 1 1

S 18q12.2 chr18:31 4 5 chr17:27 4 7 chr7:57 4 5 chr18:24 3 4

A/AB/AC 19p13.3 chr19:2 5 6 chr20:3 5 6 chr20:59 4 7 — 0 0

A/AB 19p13.2 chr19:9 10 15 chr20:9–10 9 11 chr20:54 5 7 chr9:19–20 2 9

AC 19p13.2-p13.13 chr19:11–12 24 35 chr20:12 21 32 chr20:52. 2 5 chr9:22 1 6

A/AB/AC 19p13.11-p12 chr19:19–24 38 72 chr20:20–25 35 62 — 0 0 — 0 0

chr20_random 2 2

AB 19q13.11 chr19:39–40 6 12 chr20:26–36 6 11 chr1:119–119 4 5 — 0 0

A/AB 19q13.12-q13.13 chr19:41–43 27 34 chr20:38–39 27 38 chr1:117–118 11 34 chr7:25 22 34

A/AB/Ab 19q13.31 chr19:49 20 22 chr20:46 20 20 chr1:113–113 3 13 chr7:19 11 14

A/AB/ABL 19q13.41-q13.42 chr19:57–58 39 61 chr20:54–56 37 51 chr1:106–107 8 30 chr17:18 1 1

S/SA/A/AB 19q13.43a chr19:61–62 14 17 chr20:58–59 13 18 chr1:103–104 12 18 chr7:5 7 8

S/A/AB 19q13.34b chr19:62–63 48 59 chr20:60–61 41 54 chr1:101–103 54 54 chr7:107–110 9 2

chr20_random 3 3 chrUn:17, 41 4 4

Gene types in each cluster, SCAN (S), KRAB (A, B, BL, b, or C), or combined motifs (e.g., SA = SCAN-KRAB-A).

Genome coordinates of clusters are reported with chromosome number followed by location in Mb, taken from hg17 (human), panTro1 (chimpanzee), canFam1 (dog), and mm6 (mouse) genome

sequence builds. (—) in the locus/ortholog columns signifies that no homologs were detected in that species.

Total number of predicted KRAB-ZNF protein-coding genes.

All ZNF loci including pseudogenes.

Human KRAB zinc finger family

Genome Research 671

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

served in mammals; e.g., a cluster located in HSA5q35.3 contains

six tandem genes arranged identically to orthologous genes in

human, chimp, dog, and mouse (Table 2; Supplemental Table

S6). However, several human clusters are primate specific and

most conserved gene clusters display clear evidence of lineage-

specific expansion, with different numbers of genes in each spe-

cies and relatively few orthologous groups.

To identify evolutionary relationships, we constructed a

phylogenetic tree based on KRAB-A–encoding nucleotide se-

quences (Fig. 1; Supplemental Table S2). As suggested by previous

studies focused on subsets of genes, evolutionary relatedness is

typically associated with physical proximity in this family. How-

ever, the complete family tree also shows that unrelated KRAB-

ZNF genes are physically intermixed at several clustered sites.

Therefore, although tandem in situ duplication events have rep-

resented the major mechanism of new gene creation in the

KRAB-ZNF family, distributed duplication and, possibly, post-

duplication rearrangement events have also played a prominent

role. It is uncertain at present what effect gene conversion may

have had on the evolution of these genes.

Most genes containing the KRAB-A motif also include the

KRAB-B modulator, or less common modulators KRAB-b, KRAB-

BL, or KRAB-C (Table 1). These associations appear within sepa-

rate clades in the KRAB-A–based tree, indicating that these dis-

tinct motifs arose and were expanded within specific families

(Fig. 1). Unlike genes with specific types of KRAB modulators, the

SCAN-containing genes do not group together in one evolution-

ary clade (Fig. 1, red circles). This pattern could be explained if

the SCAN-KRAB-ZNF combination is ancient, with a history of

frequent loss of one or the other effector domains during the

expansion of the gene family. This kind of history would be

consistent with the comingling of related genes with different

combinations of SCAN and KRAB effector motifs we observed in

several clusters (Table 2). However, it is also possible that the

SCAN-KRAB combination arose more than once, as has recently

been proposed (Looman et al. 2002).

Phylogenetic analyses also highlighted relatedness between

clusters and among cluster members and isolated loci distributed

at distant chromosomal sites. For example, genes from the large

19p12 cluster, which is known to be primate specific (Bellefroid

et al. 1995; Eichler et al. 1998), group together in the tree with a

clade of genes from a separate cluster located in 19q13.41. Unlike

the 19p12 cluster, this latter clade in-

cludes genes with clear mouse and dog

orthologs (Table 2, see below) and there-

fore may be related to progenitor genes

for the expanded primate group.

Relationships in certain groups

show that some distributed duplicates

may subsequently give rise to tandem

copies, suggesting one way that new lin-

eage-specific clusters may have been

seeded over evolutionary time. For ex-

ample, seven KRAB-ZNF genes and one

pseudogene sequence distributed in

HSA4, 8, 11, and 12 show >96% nucleo-

tide sequence identity over >70-kb du-

plications (ZNF705A paralogs) (high-

lighted in gray in Fig. 1; Supplemental

Table S2). The high degree of similarity

between these large segments indicates

that most of the duplication events oc-

curred ⱕ15 million years ago (Mya), and

some events may be human specific.

Most notably, two adjacent pairs of

HSA8 paralogs arose from very recent

tandem duplications (ZNF705C and

LLNL1103 at 11.9 and 12.2 Mb with

98.5% identity, and LLNL1035 and

ZNF705B, at 7.2 and 7.6 Mb with >99%

identity); although the human specific-

ity of these duplications must be veri-

fied, the chimpanzee genome contains

only one ZNF gene at each locus (chr7 at

12.7 and 7.5 Mb, respectively).

Paralogs, orthologs, and recent

primate duplications

To identify putative orthologs of human

KRAB-ZNF genes, we also generated

HMM-based gene models from the

chimpanzee, mouse, and dog genomes

Figure 1. Phylogenetic tree of human KRAB-A motifs. This neighbor-joining phylogenetic tree rep-

resents 418 human KRAB-A nucleotide sequences from KRAB-ZNF and SCAN-KRAB-ZNF genes (in-

cluding some with noncanonical structures). Gene designations are removed from this unrooted

phylogram for clarity; a list of the genes including location and aligned KRAB sequences is presented

in Supplemental Table S2. Genes from several major physical clusters are colored to show comparisons

between physical location and sequence similarity. Loci within a highlighted phylogenetic group that

do not map to the same physical cluster as related genes appear as superimposed circles in the

appropriate color (or in white for genes that do not belong to any labeled cluster). Genes that also

encode SCAN, KRAB-b, KRAB-BL, or KRAB-C motifs are indicated as follows: SCAN, red circles; KRAB-b,

orange triangles; KRAB-BL, blue squares; and KRAB-C, green diamonds. The green arrow notes the

position of the Xenopus Xfin KRAB sequence, added as a potential outgroup, although the tree is shown

unrooted.

Huntley et al.

672 Genome Research

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

and searched for reciprocal best BLAST matches between these

models and the curated human coding gene set (Supplemental

Tables S3–5). We manually inspected alignments in the variable

DNA-binding helix regions of the ZNF domains (Choo and Klug

1994; Kim and Berg 1995; Shannon et al. 2003) to confirm the

accuracy of identified orthologs (see Methods). Although the

draft status of mouse, dog, and chimp genomes makes it impos-

sible to determine the KRAB-ZNF gene repertoires of those species

completely, these data provide a clear overall view of KRAB-ZNF

gene conservation in mammals. Three hundred ninety-nine of

the 423 human KRAB-ZNF proteins detect a clear chimpanzee

ortholog, and the human and chimpanzee genes are organized

very similarly within related clustered groups (Table 2). We iden-

tified 112 genes that are conserved as 1:1 orthologs in mouse,

including 103 genes that are conserved as 1:1 reciprocal pairs in

all four mammals (Supplemental Table S6). In addition to 1:1

orthologous pairs, we found many human proteins that matched

multiple mouse and dog homologs with high similarity and, con-

versely, sets of multiple human genes that detect the same recip-

rocal best-matching gene in other species (data not shown). A

total of 226 human KRAB-ZNF protein-coding loci detected at

least one clear homolog in one or both nonprimate species

(Supplemental Table 6); the remaining 197 loci represent poten-

tial primate-specific genes (Supplemental Tables S6, S7).

Included in this set of 197 loci are KRAB-ZNF genes located

in the HSA19p12 cluster, which dates back to early primate evo-

lution and underwent a significant expansion ∼40 Mya (Belle-

froid et al. 1995; Li 1997; Eichler et al. 1998; Goodman et al.

1998). Subsequent duplications have created new HSA19 para-

logs and related genes at several distributed sites (Supplemental

Table S7; Hamilton et al. 2006). Altogether, this expanded sub-

family includes 62 of the 197 candidate primate-specific KRAB-

ZNF protein-coding genes. At least 15 genes in this subfamily are

spanned by recent segmental duplications (Bailey et al. 2001)

providing support for their recent advent. Recent segmental du-

plications also span an additional 24 nonconserved KRAB-ZNF

coding genes of other subfamilies (Supplemental Table S7). To

identify older primate duplications, we used BLAST to identify all

locus pairs with >80% sequence identity within noncoding DNA

sequences. A total of 126 genes were identified with BLAST

matches at this level or above to other human KRAB-ZNF genes,

indicating involvement in duplications that occurred in the an-

thropoid lineage (Supplemental Table S7). Combined evidence of

>80% noncoding sequence similarity, involvement in recent seg-

mental duplications, and membership in the known primate-

specific subfamily together provide a compelling argument for

the primate specificity of at least 136 protein coding loci (32% of

all human KRAB-ZNF genes).

KRAB-ZNF gene expression, cluster position, and evolutionary

history

Are clustered KRAB-ZNF genes coexpressed, or is tissue-specific

expression of tandem duplicates separately regulated and free to

diverge? To address this question, we used the Cluster software

package (Eisen et al. 1998) to analyze publicly available microar-

ray data (see Methods) from unique probes sets that could be

selected for 211 loci (Supplemental Table S8) and examined po-

tential correlations among genome location, evolutionary his-

tory, and microarray expression patterns. Although the 211

genes display diverse patterns of transcriptional activity, some

clustering of expression patterns was observed. Most notably,

expression data for 51 of the human KRAB-ZNF genes analyzed

clustered with >50% correlation into one major group, with

highest levels of expression in lymphoid tissues. A second group

of 50 genes displayed the opposite pattern, i.e., significantly re-

duced expression levels in immune-related tissues (Fig. 2; Supple-

mental Table S8). However, groups of genes with most similar

expression patterns were generally derived from different chro-

mosomal regions. In addition, a plot of Pearson’s correlation co-

efficient values for all pairwise comparisons of expression profiles

resembles a normal distribution with a median at 0.02 for this

probe set (data not shown), indicating no evidence of coregula-

tion. Although, in specific cases, recently duplicated gene pairs

have been shown to display overlapping patterns of expression

(Shannon et al. 2003), on a global scale related genes were only

infrequently grouped as best matches on the basis of expression

similarity. Reliable expression data are not available for most of

the primate-specific genes, but several sets of close relatives, such

as ZNF585A and ZNF585B (with 85% noncoding sequence iden-

tity) (Supplemental Table S7), display very different patterns of

expression (Fig. 2; Supplemental Table S8). The wide divergence

of gene expression patterns and lack of global expression corre-

lation suggest that even recent paralogs can diverge to give rise to

genes with distinct patterns of tissue-specific expression.

Discussion

We have identified and curated 423 human loci capable of en-

coding complete KRAB-ZNF proteins, including 89 novel loci;

many of the genes are alternatively spliced to encode predicted

protein isoforms with potentially very different functional prop-

erties. For example, inclusion of a KRAB-B domain has been

shown to enhance repressor activity of KRAB-A proteins (Vissing

et al. 1995), whereas the inclusion or exclusion of a SCAN do-

main may facilitate or abolish dimer formation, respectively

(Williams et al. 1999). The prevalence of these alternative tran-

scripts therefore predicts a large array of KRAB-ZNF proteins with

several different types of gene regulatory roles. Human KRAB-

ZNF proteins typically include a large number of ZNF motifs,

with a median number of 12 tandem fingers, predicting a pref-

erence for relatively long and specific DNA recognition se-

quences (Choo and Klug 1994; Berg 1997; Moore et al. 2001).

This prediction is confirmed by known binding sites of several

such “polydactyl” ZNF proteins (Schoenherr and Anderson 1995;

Zheng et al. 2000; Gebelein and Urrutia 2001; Tanaka et al.

2002). However, for certain proteins at least, only subsets of zinc

fingers are required for DNA binding (e.g., ZBRK1) (Zheng et al.

2000) and different fingers may be used to recognize targets of

distinct sequence composition and types (e.g., ZAC, Hoffmann et

al. 2003; and CTCF, Ohlsson et al. 2001). The possibility that

these polydactyl proteins might interact with multiple distinct

DNA recognition sequences through different subsets of zinc fin-

gers suggests the potential to further amplify the functional di-

versity of the KRAB-ZNF family.

As noted in previous reports, most of the 423 human KRAB-

ZNF genes reside in large familial clusters (Fig. 1, Table 2; Rous-

seau-Merck et al. 1992; Cannizzaro et al. 1993; Bellefroid et al.

1995). However, we also found a significant number of clusters

containing intermixed sets of unrelated genes and distributed

singleton copies at many sites. In some cases at least, distributed

copies have given rise to tandem duplicates, suggesting a poten-

tial means by which new clusters may be seeded at lineage-

Human KRAB zinc finger family

Genome Research 673

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

specific sites. Therefore, both tandem in situ and dispersed seg-

mental duplications have driven the expansion and the genome-

wide distribution of this gene family. In contrast to other types of

clustered families in which member genes are coexpressed in a

limited number of cell types and tissues (e.g., Mombaerts 1999;

Zhang et al. 2002; Uhrberg 2005), expression patterns of KRAB-

ZNF genes do not correlate with genomic location and neighbor-

ing genes often vary widely in transcriptional activity (Fig. 2).

These data argue strongly against coregulation of cluster neigh-

bors and indicate that tandem organization is most likely a

simple reflection of the KRAB-ZNF family’s duplicative history.

Since evolutionary relatedness correlates with genome location

in most gene clusters and even very recent paralogs can show

distinct patterns of expression, these data also argue that gene

expression patterns are free to diverge after paralog duplication.

This factor would be likely to enhance the ability of new genes to

evolve nonredundant functions (Lynch and Force 2000).

Based on comparisons between the curated human gene set

and gene models from draft mouse, dog, and chimpanzee ge-

nomes, we present a preliminary classification of human KRAB-

ZNF genes according to their degree of conservation or lineage

specificity. Although information regarding specific orthologous

relationships will change as nonhuman draft sequences are im-

proved, the overall picture of gene repertoire diversity can be

clearly discerned. Only 103 of the 423 human KRAB-ZNF genes

can be grouped in unambiguous 1:1 orthologous relationships in

primate, canine, and rodent lineages; by contrast, at least 136

loci, or nearly one-third of the total human KRAB-ZNF gene set,

are primate specific, having arisen since the emergence of Old

and New World monkeys. Since regulatory functions are known

for only a handful of KRAB-ZNF proteins, the cumulative impact

of lineage-specific gain, loss, and divergence of these genes on

primate biology remains a matter of conjecture. However, their

sheer numbers, their wide range of tissue-specific expression, and

their dynamic evolutionary history predict that KRAB-ZNF genes

have played a significant role in shaping both primate-specific

and deeply conserved traits. A more complete understanding of

the functions of the KRAB-ZNF family will be essential for deci-

phering pathways of vertebrate evolutionary diversity and for

building accurate models of gene regulation and its role in hu-

man disease.

Methods

Genome searches and initial data analysis

Human KRAB-A, KRAB-B, KRAB-b, KRAB-C, and SCAN protein

sequences were collected from RefSeq (the National Center for

Figure 2. Analysis of expression patterns for KRAB and SCAN-KRAB ZNF genes. Expression levels of members of the human KRAB and SCAN-KRAB

genes are represented by red and green boxes (denoting higher and lower expression levels, respectively). Vertical columns of boxes represent different

genes, and horizontal rows represent different tissues. We have grouped tissues into the following categories as labeled on the left of each panel: N

indicates neural; I, immune; G, glandular; M, muscle; O, other organ; and R, reproductive. In panel A, the 211 analyzed genes are arranged in

chromosomal order. Selected genomic clusters are indicated by colored boxes beneath the expression pattern. From left to right, the clusters are (as

listed in Supplemental Table S1): 16p13.3, 16p11.2a, 19p13.2, 19p13.2-p13.13, 19q13.12-q13.13, 19q13.31, 19q13.41-q13.42, 19q13.43a,

19q13.43b, 6p22.1b, 7q22.1a, 7q36.1a, and 8q24.3c. In panel B, all selected expression profiles (described in text) have been clustered based on

expression pattern similarity. Colored hash marks above the expression profiles indicate the chromosomal cluster from panel A with which each profile

corresponds.

Huntley et al.

674 Genome Research

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

Biotechnology Information mRNA reference sequence collec-

tion) (Pruitt et al. 2000) and trimmed to include only motif resi-

dues completely encoded within single exons. Zinc finger protein

sequences (X7-C-X2-C-X12-H-X3-H) from HSA19 were collected

by a simple pattern-matching script. Sequence alignments for

each motif-type were generated by using CLUSTALX (Thompson

et al. 1997) and submitted to the HMMER profile HMM matrix

building tool HMMBUILD to generate matrices (available for

download at http://znf.llnl.gov/). These matrices were used by

the HMMER search program to identify all putative motif

matches in a full six-frame translation of the hg17 genomic se-

quences.

In addition, DNA sequences from exons immediately pre-

ceding known KRAB-A exons were used to search hg17 chromo-

somal sequences by using BLAST (Altschul et al. 1990). Output

from the HMMER and BLAST searches was compared with the

genomic coordinates of publicly available gene models from

RefSeq, UCSC Known (known protein-coding genes based on

protein data from UniProt, i.e., SWISS-PROT and TrEMBL, and

mRNA data from RefSeq), and MGC (Mammalian Gene Collec-

tion gene models) (Strausberg et al. 1999) to identify previously

characterized loci. The search results were also arranged in chro-

mosomal order and grouped based on proximity and orientation

into putative loci.

The human HMM matrices were also used to search the

chimp (panTro1), mouse (mm6), and dog (canFam1) six-frame

genome translations, and putative loci were generated based on

proximity and orientation as above. Crude protein sequences

for these nonhuman loci were generated by extending from mo-

tif coordinates N- and C-terminally until a translational stop

signal was encountered, eliminating overlapping sequences

from adjacent motifs, and joining all collected sequences for

each locus.

Gene annotation and database curation

Gene model structures were manipulated by using a modified

version of APOLLO (http://www.fruitfly.org/annot/apollo/) dis-

playing publicly available mRNA, EST, and other evidence as

well as HMM-identified motifs. Public gene models were revised

to coincide with RNA evidence. New models were created based

on RNA evidence where available but, in some cases, were gen-

erated de novo so as to include an open reading frame containing

the motifs and maintaining canonical intron splice sites (GT-AG,

AT-AG, and GC-AG). Loci were classified as pseudogenes if no

gene model could be made that could produce a functional pro-

tein. Pseudogenes with limited stop codons or insertions/

deletions were re-evaluated for functionality by examining the

putative open reading frames of any RNA sequences associated

with the locus and identifying discrepancies with the genomic

sequence.

Adjacent genes were considered “clustered” if the intergenic

sequence separating two KRAB-ZNF genes was <200 kb, even in

cases where unrelated genes were found between the ZNF loci (a

situation only rarely encountered). This distance cutoff was se-

lected based on a distribution of intergenic distances between

KRAB-ZNF genes in the annotation database (for additional de-

tails, see legend of Supplemental Fig. S1). Manual inspection of

the clustering confirmed that the 200-kb criterion resulted in

coclustering of all major groups of neighboring, related genes.

Only two pairs of paralogs that could be considered to form fa-

milial clusters were omitted by the 200-kb cutoff; these genes—

LLNL1035 and ZNF705B, and ZNF705C and LLNL1103—derived

from unusually large duplicons were counted as clusters in the

final analysis.

Comparative genomic analyses

Protein sequences from all human KRAB and SCAN-KRAB loci

were BLASTed against the full set of human ZNF protein se-

quences and against draft protein sequences predicted from

chimp, mouse, or dog ZNF loci. The best match (with expectation

value <10

ⳮ30

and >30% identity) was then BLASTed back against

the human protein set; if this second BLAST returned the initial

search query, the pair was considered a reciprocal best match.

Data were collected for the best six matches for each protein, and

reciprocal matches were flagged. The data were then manually

checked to identify and correct anomalous results by inspection

of the variable helix regions of aligned zinc fingers. Most inac-

curate identifications were due to sequencing errors, e.g., those

predicting truncated proteins in the draft genome searched; this

issue was most acute where very closely related paralogs exist in

both species. In cases where a true ortholog is missing from a

draft genome sequence, the most similar sequenced paralogous

gene will be selected by this method. Chromosomal and relative

positions of reciprocal matches were compared both within the

human genome and across the four species to evaluate lineage-

specific cluster evolution.

Evolutionary analysis

A tree of phylogenetic relationships was generated by using

KRAB-A motifs from KRAB-ZNF and KRAB-SCAN-ZNF loci, in-

cluding several genes with noncanonical structures. Several loci

contained two KRAB-A motifs; in these cases both KRAB-A motifs

were used. KRAB-A nucleotide sequences were retrieved from the

catalog Web site, and alignments of the sequences were made by

using CLUSTALX 1.81 (Thompson et al. 1997). The alignment

was manually checked by using SeAl (Rambaut 1996). The PAUP

4.0b10 package (Swofford 2002) was used to generate trees by

using mean character differences and the neighbor-joining (NJ)

method (Saitou and Nei 1987). A Xenopus KRAB-A (Xfin) se-

quence was added as a potential outgroup.

Selection of microarray probe sets and expression clustering

Affymetrix-based GNF Atlas expression data pre-analyzed using

the MAS5 algorithm (Su et al. 2004) was obtained from the UCSC

Genome Bioinformatics Web site (http://genome.ucsc.edu/, Kent

et al. 2002), and GNF1H and U133A probe sets relevant to genes

in our data set were selected. We included data only from probes

that could reasonably be expected to detect expression of a single

gene; due to the high degree of sequence similarity between ZNF

paralogs, these represented a very small subset of the probe sets

designed against KRAB-ZNF genes. To identify relatively unique

probe sets, the individual sequences for all probe sets from both

chip designs were BLASTed against all catalog transcript se-

quences (including noncanonical genes). A pool of gene-specific

probe sets was then identified based on two criteria: (1) all se-

quences within a given probe set perfectly aligned to some por-

tion of a single target locus; (2) the ratio of the number of target-

specific alignments over the total number of alignments to any

locus (i.e., target and nontarget loci to which probes had ⱖ20/25

matches) was >0.8. This second requirement eliminates probe

sets that interrogate additional target(s) well or numerous targets

poorly. In instances where multiple candidate probe sets were

identified for a single gene, a representative probe set was manu-

ally chosen, selecting probes aligning with 3⬘ regions of the tran-

scripts wherever possible.

By use of Cluster v. 2.11 (Eisen et al. 1998), the expression

profiles for the selected probe sets were hierarchically clustered

using an uncentered correlation metric. The resulting cdt files

were visualized by using TreeView v. 1.60 (Eisen et al. 1998) to

Human KRAB zinc finger family

Genome Research 675

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

generate Figure 2B. Profiles were also manually arranged in chro-

mosomal order and visualized in TreeView to generate Figure 2A.

These expression profiles were also used to analyze possible glo-

bal coregulation by using Pearson’s correlations coefficient.

Acknowledgments

We thank Ivan Ovcharenko for advice on programming and ge-

nome searching methods, and David Goodstein and Astrid Terry

at the Joint Genome Institute for advice on Apollo and gene

annotation. We also thank Colleen Elso, Jutta Kollet, Jason Ray-

mond, and Alice Yamada for critical reviews of the manuscript

and Web site. This work was performed under the auspices of the

U.S. Department of Energy by the University of California,

Lawrence Livermore National Laboratory (LLNL) under contract

no. W-7405-Eng-48. The project (04-ERD-084) was funded by the

Laboratory Directed Research and Development Program at

LLNL.

References

Abrink, M., Ortiz, J.A., Mark, C., Sanchez, C., Looman, C., Hellman, L.,

Chambon, P., and Losson, R. 2001. Conserved interaction between

distinct Kruppel-associated box domains and the transcriptional

intermediary factor 1␤. Proc. Natl. Acad. Sci. 98: 1422–1426.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. 1990.

Basic local alignment search tool. J. Mol. Biol. 215: 403–410.

Ayyanathan, K., Lechner, M.S., Bell, P., Maul, G.G., Schultz, D.C.,

Yamada, Y., Tanaka, K., Torigoe, K., and Rauscher III, F.J. 2003.

Regulated recruitment of HP1 to a euchromatic gene induces

mitotically heritable, epigenetic gene silencing: A mammalian cell

culture model of gene variegation. Genes & Dev. 17: 1855–1869.

Bailey, J.A., Yavor, A.M., Massa, H.F., Trask, B.J., and Eichler, E.E. 2001.

Segmental duplications: Organization and impact within the current

human genome project assembly. Genome Res. 11: 1005–1017.

Bellefroid, E.J., Poncelet, D.A., Lecocq, P.J., Revelant, O., and Martial,

J.A. 1991. The evolutionarily conserved Kruppel-associated box

domain defines a subfamily of eukaryotic multifingered proteins.

Proc. Natl. Acad. Sci. 88: 3608–3612.

Bellefroid, E.J., Marine, J.C., Ried, T., Lecocq, P.J., Riviere, M., Amemiya,

C., Poncelet, D.A., Coulie, P.G., de Jong, P., Szpirer, C., et al. 1993.

Clustered organization of homologous KRAB zinc-finger genes with

enhanced expression in human T lymphoid cells. EMBO J.

12: 1363–1374.

Bellefroid, E.J., Marine, J.C., Matera, A.G., Bourguignon, C., Desai, T.,

Healy, K.C., Bray-Ward, P., Martial, J.A., Ihle, J.N., and Ward, D.C.

1995. Emergence of the ZNF91 Kruppel-associated box-containing

zinc finger gene family in the last common ancestor of

anthropoidea. Proc. Natl. Acad. Sci. 92: 10757–10761.

Berg, J.M. 1997. Letting your fingers do the walking. Nat. Biotechnol.

15: 323.

Cannizzaro, L.A., Aronson, M.M., and Thiesen, H.J. 1993. Human zinc

finger gene ZNF23 (Kox16) maps to a zinc finger gene cluster on

chromosome 16q22, and ZNF32 (Kox30) to chromosome region

10q23-q24. Hum. Genet. 91: 383–385.

Choo, Y. and Klug, A. 1994. Selection of DNA binding sites for zinc

fingers using rationally randomized DNA reveals coded interactions.

Proc. Natl. Acad. Sci. 91: 11168–11172.

Chung, H.R., Schafer, U., Jackle, H., and Bohm, S. 2002. Genomic

expansion and clustering of ZAD-containing C2H2 zinc-finger genes

in Drosophila. EMBO Rep. 3: 1158–1162.

Collins, T., Stone, J.R., and Williams, A.J. 2001. All in the family: The

BTB/POZ, KRAB, and SCAN domains. Mol. Cell. Biol. 21: 3609–3615.

Eichler, E.E., Hoffman, S.M., Adamson, A.A., Gordon, L.A., McCready,

P., Lamerdin, J.E., and Mohrenweiser, H.W. 1998. Complex

␤-satellite repeat structures and the expansion of the zinc finger

gene cluster in 19p12. Genome Res. 8: 791–808.

Eisen, M.B., Spellman, P.T., Brown, P.O., and Botstein, D. 1998. Cluster

analysis and display of genome-wide expression patterns. Proc. Natl.

Acad. Sci. 95: 14863–14868.

Friedman, J.R., Fredericks, W.J., Jensen, D.E., Speicher, D.W., Huang,

X.P., Neilson, E.G., and Rauscher III, F.J. 1996. KAP-1, a novel

corepressor for the highly conserved KRAB repression domain. Genes

& Dev. 10: 2067–2078.

Gebelein, B. and Urrutia, R. 2001. Sequence-specific transcriptional

repression by KS1, a multiple-zinc-finger-Kruppel-associated box

protein. Mol. Cell. Biol. 21: 928–939.

Goodman, M., Porter, C.A., Czelusniak, J., Page, S.L., Schneider, H.,

Shoshani, J., Gunnell, G., and Groves, C.P. 1998. Toward a

phylogenetic classification of primates based on DNA evidence

complemented by fossil evidence. Mol. Phylogenet. Evol. 9: 585–598.

Hamilton, A.T., Huntley, S., Kim, J., Branscomb, E., and Stubbs, L. 2003.

Lineage-specific expansion of KRAB zinc-finger transcription factor

genes: Implications for the evolution of vertebrate regulatory

networks. Cold Spring Harb. Symp. Quant. Biol. 68: 131–140.

Hamilton, A.T., Huntley, S., Tran-Gyamfi, M., Baggott, D.M., Gordon,

L., and Stubbs, L. 2006. Evolutionary expansion and divergence in

the ZNF91 subfamily of primate-specific zinc finger genes. Genome

Res. (this issue).

Hoffmann, A., Ciani, E., Boeckardt, J., Holsboer, F., Journot, L., and

Spengler, D. 2003. Transcriptional activities of the zinc finger

protein Zac are differentially controlled by DNA binding. Mol. Cell.

Biol. 23: 988–1003.

Huntley, S., Hamilton, A., Kim, J., Branscomb, E., and Stubbs, L.

Tandem gene family expansion and genomic diversity. In

Comparative genomics: A guide to the analysis of eukaryotic genomes (ed.

M.D. Adams). Humana Press, New York, (in press).

Kent, W.J., Sugnet, C.W., Furey, T.S., Roskin, K.M., Pringle, T.H., Zahler,

A.M., and Haussler, D. 2002. The human genome browser at UCSC.

Genome Res. 12: 996–1006.

Kim, C.A. and Berg, J.M. 1995. Serine at position 2 in the DNA

recognition helix of a Cys2-His2 zinc finger peptide is not, in

general, responsible for base recognition. J. Mol. Biol. 252: 1–5.

Knochel, W., Poting, A., Koster, M., el Baradi, T., Nietfeld, W.,

Bouwmeester, T., and Pieler, T. 1989. Evolutionary conserved

modules associated with zinc fingers in Xenopus laevis. Proc. Natl.

Acad. Sci. 86: 6097–6100.

Krebs, C.J., Larkins, L.K., Khan, S.M., and Robins, D.M. 2005. Expansion

and diversification of KRAB zinc-finger genes within a cluster

including regulator of sex-limitation 1 and 2. Genomics 85: 752–761.

Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C.,

Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al.

2001. Initial sequencing and analysis of the human genome. Nature

409: 860–921.

Li, W.H. 1997. Molecular evolution. Sinauer Associates, Sunderland, MA.

Looman, C., Abrink, M., Mark, C., and Hellman, L. 2002. KRAB zinc

finger proteins: An analysis of the molecular mechanisms governing

their increase in numbers and complexity during evolution. Mol.

Biol. Evol. 19: 2118–2130.

Looman, C., Hellman, L., and Abrink, M. 2004. A novel

Kruppel-associated box identified in a panel of mammalian zinc

finger proteins. Mamm. Genome 15: 35–40.

Lynch, M. and Force, A. 2000. The probability of duplicate gene

preservation by subfunctionalization. Genetics 154: 459–473.

Margolin, J.F., Friedman, J.R., Meyer, W.K., Vissing, H., Thiesen, H.J.,

and Rauscher III, F.J. 1994. Kruppel-associated boxes are potent

transcriptional repression domains. Proc. Natl. Acad. Sci.

91: 4509–4513.

Mark, C., Abrink, M., and Hellman, L. 1999. Comparative analysis of

KRAB zinc finger proteins in rodents and man: Evidence for several

evolutionarily distinct subfamilies of KRAB zinc finger genes. DNA

Cell Biol. 18: 381–396.

Messina, D.N., Glasscock, J., Gish, W., and Lovett, M. 2004. An

ORFeome-based analysis of human transcription factor genes and

the construction of a microarray to interrogate their expression.

Genome Res. 14: 2041–2047.

Miller, J.C. and Pabo, C.O. 2001. Rearrangement of side-chains in a

Zif268 mutant highlights the complexities of zinc finger–DNA

recognition. J. Mol. Biol. 313: 309–315.

Mombaerts, P. 1999. Odorant receptor genes in humans. Curr. Opin.

Genet. Dev. 9: 315–320.

Moore, M., Klug, A., and Choo, Y. 2001. Improved DNA binding

specificity from polyzinc finger peptides by using strings of

two-finger units. Proc. Natl. Acad. Sci. 98: 1437–1441.

Oh, H.J., Li, Y., and Lau, Y.F. 2005. Sry associates with the

heterochromatin protein 1 complex by interacting with a KRAB

domain protein. Biol. Reprod. 72: 407–415.

Ohlsson, R., Renkawitz, R., and Lobanenkov, V. 2001. CTCF is a

uniquely versatile transcription regulator linked to epigenetics and

disease. Trends Genet. 17: 520–527.

Pengue, G. and Lania, L. 1996. Kruppel-associated box-mediated

repression of RNA polymerase II promoters is influenced by the

arrangement of basal promoter elements. Proc. Natl. Acad. Sci.

93: 1015–1020.

Pruitt, K.D., Katz, K.S., Sicotte, H., and Maglott, D.R. 2000. Introducing

RefSeq and LocusLink: Curated human genome resources at the

Huntley et al.

676 Genome Research

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

NCBI. Trends Genet. 16: 44–47.

Rambaut, A. 1996. Se-Al: Sequence Alignment Editor.

http://iubio.bio.indiana.edu/soft/iubionew/molbio/dna/analysis/

Pist/main.html

Rousseau-Merck, M.F., Tunnacliffe, A., Berger, R., Ponder, B.A., and

Thiesen, H.J. 1992. A cluster of expressed zinc finger protein genes

in the pericentromeric region of human chromosome 10. Genomics

13: 845–848.

Saitou, N. and Nei, M. 1987. The neighbor-joining method: A new

method for reconstructing phylogenetic trees. Mol. Biol. Evol.

4: 406–425.

Sander, T.L. and Morris, J.F. 2002. Characterization of the SCAN box

encoding RAZ1 gene: Analysis of cDNA transcripts, expression, and

cellular localization. Gene 296: 53–64.

Sander, T.L., Stringer, K.F., Maki, J.L., Szauter, P., Stone, J.R., and

Collins, T. 2003. The SCAN domain defines a large family of zinc

finger transcription factors. Gene 310: 29–38.

Schmidt, D. and Durrett, R. 2004. Adaptive evolution drives the

diversification of zinc-finger binding domains. Mol. Biol. Evol.

21: 2326–2339.

Schoenherr, C.J. and Anderson, D.J. 1995. The neuron-restrictive

silencer factor (NRSF): A coordinate repressor of multiple

neuron-specific genes. Science 267: 1360–1363.

Shannon, M. and Stubbs, L. 1998. Analysis of homologous

XRCC1-linked zinc-finger gene families in human and mouse:

Evidence for orthologous genes. Genomics 49: 112–121.

Shannon, M., Hamilton, A.T., Gordon, L., Branscomb, E., and Stubbs, L.

2003. Differential expansion of zinc-finger transcription factor loci

in homologous human and mouse gene clusters. Genome Res.

13: 1097–1110.

Strausberg, R.L., Feingold, E.A., Klausner, R.D., and Collins, F.S. 1999.

The mammalian gene collection. Science 286: 455–457.

Su, A.I., Wiltshire, T., Batalov, S., Lapp, H., Ching, K., Block, D., Zhang,

J., Soden, R., Hayakawa, M., Kreiman, G., et al. 2004. A gene atlas of

the mouse and human protein-encoding transcriptomes. Proc. Natl.

Acad. Sci. 101: 6062–6067.

Swofford, D.L. 2002. PAUP: Phylogenetic analysis using parsimony. Sinauer

Associates, Sunderland, MA.

Tanaka, K., Tsumaki, N., Kozak, C.A., Matsumoto, Y., Nakatani, F.,

Iwamoto, Y., and Yamada, Y. 2002. A Kruppel-associated box-zinc

finger protein, NT2, represses cell-type-specific promoter activity of

the ␣2(XI) collagen gene. Mol. Cell. Biol. 22: 4256–4267.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., and Higgins,

D.G. 1997. The CLUSTAL_X windows interface: Flexible strategies

for multiple sequence alignment aided by quality analysis tools.

Nucleic Acids Res. 25: 4876–4882.

Uhrberg, M. 2005. The KIR gene family: Life in the fast lane of

evolution. Eur. J. Immunol. 35: 10–15.

Venter, J.C., Adams, M.D., Myers, E.W., Li, P.W., Mural, R.J., Sutton,

G.G., Smith, H.O., Yandell, M., Evans, C.A., Holt, R.A., et al. 2001.

The sequence of the human genome. Science 291: 1304–1351.

Vissing, H., Meyer, W.K., Aagaard, L., Tommerup, N., and Thiesen, H.J.

1995. Repression of transcriptional activity by heterologous KRAB

domains present in zinc finger proteins. FEBS Lett. 369: 153–157.

Williams, A.J., Blacklow, S.C., and Collins, T. 1999. The zinc

finger-associated SCAN box is a conserved oligomerization domain.

Mol. Cell. Biol. 19: 8526–8535.

Wu, Y., Yu, L., Bi, G., Luo, K., Zhou, G., and Zhao, S. 2003.

Identification and characterization of two novel human SCAN

domain-containing zinc finger genes ZNF396 and ZNF397. Gene

310: 193–201.

Zhang, H.B., Liu, D.P., and Liang, C.C. 2002. The control of expression

of the ␣-globin gene cluster. Int. J. Hematol. 76: 420–426.

Zheng, L., Pan, H., Li, S., Flesken-Nikitin, A., Chen, P.L., Boyer, T.G.,

and Lee, W.H. 2000. Sequence-specific transcriptional corepressor

function for BRCA1 through a novel zinc finger protein, ZBRK1. Mol.

Cell 6: 757–768.

Received October 21, 2005; accepted in revised form March 6, 2006.

Human KRAB zinc finger family

Genome Research 677

www.genome.org

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from

10.1101/gr.4842106Access the most recent version at doi:

2006 16: 669-677 Genome Res.

Stuart Huntley, Daniel M. Baggott, Aaron T. Hamilton, et al.

transcriptional repressors

genes: Insights into the evolutionary history of a large family of

A comprehensive catalog of human KRAB-associated zinc finger

Material

Supplemental

http://genome.cshlp.org/content/suppl/2006/04/10/gr.4842106.DC1

References

http://genome.cshlp.org/content/16/5/669.full.html#ref-list-1

This article cites 56 articles, 26 of which can be accessed free at:

License

Service

Email Alerting

click here.top right corner of the article or

Receive free email alerts when new articles cite this article - sign up in the box at the

https://genome.cshlp.org/subscriptions

go to: Genome Research To subscribe to

Cold Spring Harbor Laboratory Press

Cold Spring Harbor Laboratory Press on September 10, 2024 - Published by genome.cshlp.orgDownloaded from