Tải bản đầy đủ - 0 (trang)
2 Targeting by Ig Promoters – Are High Levels ofTranscription All There is to It?

2 Targeting by Ig Promoters – Are High Levels ofTranscription All There is to It?

Tải bản đầy đủ - 0trang


DNA Deamination and the Immune System

bubble created by the DNA polymerase is responsible for this effect.

This requirement for transcription holds true when ectopicallyexpressing AID in E. coli and yeast, and assessing the resulting mutator

phenotypes (Sohail et al., 2003; Ramiro et al., 2003; Petersen-Mahrt

et al., 2002; Poltoratsky et al., 2004). These findings are consistent with

the fact that B cells, by definition, transcribe their rearranged Ig genes

and express/secrete the encoded Ig proteins. SHM is strictly dependent

on Ig gene transcription (Fukita et al., 1998), and CSR requires the

generation of sterile (non-coding) transcripts driven by promoters

upstream of switch repeat regions (Stavnezer, 2000). Furthermore, short

single-stranded regions are found in V region DNA, coincident with

transcription in B cells (Ronai et al., 2007). These requirements for

transcription, and the observation that Ig genes are transcribed at

particularly high levels in B cells, led to the emergence of the idea that

high levels of transcription (and not cis-regulatory elements) might be all

that is required for targeting of AID-mediated sequence diversification to

a given gene. This model is consistent with the observations of AIDdependent mutagenesis in various stably integrated expression cassettes

lacking any Ig gene sequences in hybridoma lines, Ramos cells,

fibroblasts and CHO cells (Martin and Scharff, 2002; Martin et al., 2002;

Yoshikawa et al., 2002). It is important to note, however, that these

effects required the ectopic overexpression of AID in these cells,

suggesting that high levels of AID might override the action of cis-acting

targeting elements. Similarly, in non-B cells the targeting machinery

might simply be absent or inactive which would also explain why the

ubiquitous expression of AID in transgenic mice led to non-B cell tumors

without evidence for increased SHM rates in B cells (Okazaki et al., 2003).

3.2.1 Genome-wide SHM

One of the strongest challenges to the model of specific cis-acting

targeting elements came from a study performed by the Wabl laboratory

that used a retrovirus containing a GFP reporter gene to test the

mutability of numerous randomly selected gene loci (Wang et al., 2004).

Surprisingly, the reporter got mutated, as measured by the emergence of

GFP+ cells, in an AID-dependent fashion but independent of where in the

Cis-Regulatory Elements that Target AID to Immunoglobulin Loci


genome the reporter was integrated. As the transcription of the reporter

gene was driven by the viral LTR promoter, it is consistent with the

concept that transcription is all that is required for SHM to occur. A

challenge to this interpretation is the very low levels of mutations in this

system compared to SHM in activated B cells. Consistent with the low

level of mutation, neither of the B cell lines (70Z/3 and 18-81) is of

germinal center origin and hence might lack the active targeting

machinery. Lastly, a large-scale study in which 118 expressed genes

were sequenced from Peyer’s patch B cells revealed that one-quarter of

genes analyzed was susceptible to SHM (Liu et al., 2008). Importantly,

the mutation loads were again very low compared to those found in the

Ig genes: in some of the genes mutations could only be detected when

DNA repair pathways were disrupted. Overall, while all these reports

strongly suggest that non-Ig genes can be targets of SHM, it remains true

that the high levels of AID-mediated sequence diversification observed

in bone fide SHM of V region genes, are exclusive to Ig loci. It is

conceivable, however, that genes mutated by mistargeting, contain

“attenuated” or partial targeting elements, i.e. only some of the important

binding sites are present, or their order/orientation is suboptimal. The

identification of the molecular mechanism of targeting will be critical to

address this hypothesis.

3.2.2 Targeting of SHM by promoters

While it is widely accepted that a transcriptional promoter is essential for

all AID-mediated sequence diversification processes, the question

whether the very nature of the promoter is important remains open.

Some studies show that any promoter is sufficient, but others observe

that the ability to drive transcription and the permissiveness for SHM do

not correlate. The first notable observation was made in knockout/knock-in mice in which the endogenous VH promoter (which is

PolII-dependent) was either deleted or replaced with a PolI-dependent

promoter (Fukita et al., 1998). While the presence of a promoter was

required for SHM, even the PolI promoter was unexpectedly able to

support it. There was some indication, however, that a cryptic PolIIdependent promoter upstream of the rearranged IgH gene was driving the


DNA Deamination and the Immune System

SHM process in this instance. Similarly, replacing the Igκ promoter in

Igκ transgenes with a heterologous human β-globin promoter, did not

affect SHM of the transgene (Betz et al., 1994). But given concerns

about the robustness of this model system (discussed below), the

relevance of this finding for the endogenous Ig gene loci is unclear.

Recent promoter-swapping experiments in the chicken DT40 B cell line

(this system is discussed in Section 3.2) revealed that the nature of the

promoter itself can drastically alter the frequency of AID-dependent

sequence diversification (Yang et al., 2006). When the constitutive

EF1α and β-actin promoters were used to replace the endogenous IgL

promoter, both were capable of driving high levels of transcription but

only the β-actin promoter supported relatively normal levels of AIDmediated sequence diversification. Similar observations were made

when the EF1α promoter was replacing the IgL promoter in randomlyintegrated reporter constructs containing targeting elements (Kim and

Tian, 2009). In contrast, the very same EF1α-driven reporter (and a

matching β-actin promoter driven reporter) was getting mutated when

integrated in the IgL locus (Kim and Tian, 2009). Note that in this case

the endogenous IgL promoter was left intact and might have contributed

to targeting. Lastly, the strong Rous Sarcoma Virus (RSV) promoter was

sufficient to support SHM of a GFP reporter gene even in the absence of

the endogenous IgL promoter (Blagodatski et al., 2009). Overall, these

reports suggest that Ig promoters do play an important role in targeting

SHM beyond solely driving transcription of the Ig genes. Although Ig

promoters are not essential, they do increase the frequency of SHM.

Such function is consistent with the model that looping brings cis-acting

targeting elements in close proximity to the promoters, facilitating the

recruitment of AID and/or the transfer of the AID to the transcription

machinery that is poised to traverse along the Ig gene (Fig. 3.1). The

subsequent movement of the PolII complex along the gene is thought to

deliver AID to its cognate DNA substrate, the coding region of the Ig

genes for SHM and the switch region for CSR, respectively. Some nonIg promoters might be permissive for functional interactions with the cisacting targeting elements and hence SHM (or CSR), and some might lack

binding sites for trans-acting factors critical for looping, and thus do not

drive these processes efficiently.

Cis-Regulatory Elements that Target AID to Immunoglobulin Loci


Figure 3.1. Targeting of SHM and CSR. A schematic representation of the targeting

element model for restricting AID action to Ig gene loci is shown. (A) SHM, here

depicted for an Ig light chain locus, requires the concerted interaction of the

transcriptional enhancer (E) and the targeting element (TE) with the Ig promoter (P) to

establish a platform to which RNA polymerase II holoenzyme complexes and AID are

recruited. AID then trails with the polymerase complex along the Ig gene and converts

Cs to Us in the transiently-formed single-stranded DNA within the VJ exon. The U:G

mismatches are subsequently fixed by direct replication and/or error-prone DNA repair to

give rise to the SHM end product, a muted V region. (B) CSR requires the concerted

interaction of the transcriptional enhancer (E), the targeting element (TE) with the

intronic promoters (here schematically shown for Iµ and Iγ), which recruits RNA

polymerase II and AID to both promoters, and also induces synapsis of the switch repeat

regions (Sµ and Sγ, respectively). A high frequency of DNA deamination events within

single-stranded DNA formed by transcription of the repeats, leads to DNA double-strand

breaks, and ends up with the final CSR product. Note that although transcriptional

enhancer and targeting elements are shown as distinct entities, they are likely to show

physical overlap in the genome. In addition, the exact number of cis-regulatory elements

required is unknown, and for simplicity a single element model is shown.


DNA Deamination and the Immune System

3.3 SHM Targeting Elements in Ig Light Chain Loci

3.3.1 The murine Ig light chain loci

The starting point in the search for cis-acting targeting elements in the

murine Igκ was the observation that transgenes resembling a rearranged

Igκ locus were subject to SHM in vivo (O'Brien et al., 1987; Sharpe

et al., 1991). Similar effects were also observed using Igλ transgenes

(Klotz and Storb, 1996; Kong et al., 1998). As the murine Igλ locus

is far less well characterized in terms of cis-regulatory elements,

subsequent studies focused almost exclusively on the Igκ gene locus. An

important feature of the Igκ transgenic constructs was that they were

designed as passenger transgenes, i.e. they do not confer expression of a

functional Igκ chain. Thus such transgenes do not interfere with B cell

development, are not subject to selective forces during germinal center

responses and are thought to solely act as reporters for SHM.

A key observation from the analysis of mice harboring such

transgenes was that heterologous non-Ig sequences can be mutated by

SHM within the transgenes (Yelamos et al., 1995). This is in striking

contrast to CSR where the nature of the sequence on which AID acts is

an essential component for the process to occur (Zarrin et al., 2004).

Thus, the search for cis-acting targeting elements focused on the noncoding sequences within the Igκ locus. As transcription and SHM

are intimately linked, the importance of two well-characterized

transcriptional control elements for SHM was tested. Transgenes lacking

the intronic κ enhancer (iEκ; located between Jκ5 and Cκ), and its

flanking matrix attachment region (MAR) showed no detectable SHM,

while transcription was unaffected (Betz et al., 1994). In contrast, Igκ

transgenes lacking the 3′ κ enhancer (3′κE) showed a dramatic reduction

in both transcription and SHM (Betz et al., 1994). These observations

suggested that cis-acting targeting elements reside in the iEκ/MAR,

while the evidence for such sequences in the 3′κE was inconclusive.

Subsequent attempts using this model system to identify the minimal

targeting elements for SHM within the iEκ/MAR and the 3′κE by

systematic deletion strategies turned out to reveal uninformative results

(Goyenechea et al., 1997; Klix et al., 1998). The problems were likely to

Cis-Regulatory Elements that Target AID to Immunoglobulin Loci


be of a technical nature, as the behavior of transgenic constructs is

frequently strongly dependent on their integration site. In particular,

deleting seemingly non-functional “junk” DNA exposes minimal

regulatory elements to the influence of their chromosomal environment,

and can lead to variegated results. There is anecdotal evidence that a

subset of Igκ transgenes containing the full set of regulatory elements do

not undergo SHM while still being highly transcribed.

Gene-targeting approaches were used, to avoid the intrinsic

challenges encountered when using transgenic constructs to study cisregulatory elements. In contrast to the observations in transgenes, the

iEκ element was found to be dispensable for SHM when it was deleted

from the endogenous Igκ locus (Inlay et al., 2006). Similarly, the 3′κE

enhancer was not essential for SHM, but its deletion led to reduced

mutation frequencies that are in part explained by reduced Igκ transcript

levels (van der Stoep et al., 1998; Inlay et al., 2006). Lastly, the deletion

of the distal enhancer (Ed) led to a similar reduction of both SHM and

transcription (Xiang and Garrard, 2008). These findings raise the

possibility of redundancy between these transcriptional control elements

with respect to SHM targeting. The combined deletion of iEκ and 3′κE

results in an early block of B cell development caused by a lack of

transcription and hence V(D)J recombination (Inlay et al., 2002), and

thus the question of redundancy still awaits to be addressed. Two

alternative approaches emerge to bypass the importance of the enhancer

elements for transcription during B cell development: (1) the knock-in

of exogenous, constitutively active non-Ig promoters allowing for

transcription even in the absences of Ig enhancers; and (2) the knock-in

of exogenous, constitutively-active non-Ig enhancers. The former might

suffer from problems discussed above (non-Ig promoters support only

reduced levels of SHM), while the latter might not be sufficient to rescue

early B cell development, as is the case for the SV40 enhancer in the

context of the IgH locus (Kuzin et al., 2008). In this context, it is also

worth noting that the question whether the Igκ coding region is truly

irrelevant, has not been formally addressed by gene-targeting thus far.

Overall, the studies of transcriptional enhancers within the murine

Igκ locus and their role in SHM, strongly suggest that cis-acting

targeting elements do exist. They are likely to be linked to, but not


DNA Deamination and the Immune System

necessarily identical with, these transcriptional control elements.

Furthermore, as these elements are sensitive with respect to their

chromosomal environment, it is obvious that gene-targeting strategies or

the use of BAC transgenes (being less susceptible to integration site

effects, as discussed below for the IgH locus), are the methods of choice

to identify these elements.

3.3.2 The chicken IgL locus

The chicken DT40 B cell line is currently considered one of the most

robust model systems to address the issue of targeting of SHM. Under

standard culture conditions, DT40 cells express AID and continuously

undergo SHM and Ig gene conversion (GCV) of the IgH and the single

avian IgL gene (Arakawa and Buerstedde, 2004). GCV is closely related

to SHM, and thought to only differ in the repair phase of the process,

using upstream pseudo-V (ψV) genes as sequence donors for homologybased repair, leading to multiple nucleotide changes per event. The

mechanisms that govern targeting of SHM and GCV to Ig loci appear

identical, as deletion of all ψV genes shifts the cells to exclusively

performing SHM (Arakawa et al., 2004), albeit with a mutation pattern

distinct from that observed in humans and mice. Thus the terms SHMand GCV-targeting are used interchangeably within this section. For the

purpose of studying targeting elements, DT40 cells have a key advantage

over many other cell line models, namely that standard gene-targeting

strategies can be used to manipulate their genome (Arakawa and

Buerstedde, 2006). Lastly, the availability of the chicken genome

sequence set the stage for the analysis of cis-acting DNA sequences in

the chicken IgL locus. The genomic assembly of the chicken IgH locus

is incomplete in its current state, and thus all studies are focused on the

IgL locus, which is more closely related to mammalian Igλ than Igκ.

Three groups employed systematic deletion approaches to identify the

cis-acting sequences that are required for targeting AID-mediated DNA

diversification processes to the IgL gene (Kothapalli et al., 2008;

Blagodatski et al., 2009; Kim and Tian, 2009). Although all three

studies use slightly different strategies, they converge on the same region

Cis-Regulatory Elements that Target AID to Immunoglobulin Loci


of the IgL gene locus located downstream of the CL constant region

exons (Fig. 3.2). The first report of a targeting element by the Fugmann

laboratory used the endogenous Ig promoter-driven IgL gene (supported

by a SV40 enhancer when endogenous enhancer elements were deleted)

as a read-out, and deleted an increasing amount of non-coding DNA

from the IgL locus (Kothapalli et al., 2008). This led to the identification

of the 3′ regulatory region (3′RR) containing a transcriptional enhancer

and a targeting element for SHM. Subsequently, the targeting function

was assigned to a 1.3kb subfragment (now named mutational enhancer

element, MEE) of the 3′RR, that is essential for SHM but dispensable for

transcription in an otherwise unaltered IgL gene locus (N. Kothapalli and

S.D. Fugmann, unpublished data; Fig. 3.2). The Buerstedde laboratory

used a RSV promoter-driven GFP reporter to assess SHM in the context

of an IgL locus in which all pseudo-V elements had been deleted, and

identified a bipartite targeting element that was named diversification

activator (DIVAC; Blagodatski et al., 2009). Importantly, a DNA

fragment containing the DIVAC together with the RSV promoter-driven

GFP reporter was sufficient to support AID-mediated sequence

diversification when inserted in non-Ig genes. Whether the genes in

which these constructs were integrated, were also rendered targets of

SHM has not been analyzed thus far. Lastly, the Tian laboratory used a

puromycin reporter to assess gene conversion activities, and identified a

region “A” that is critical for targeting in the IgL locus, and sufficient for

targeting when integrated randomly in the genome, albeit at ten-fold

lower frequencies than observed in the IgL locus (Kim and Tian, 2009).

This study again focused solely on the reporter gene read-out, and did

not report the effects on the respective endogenous IgL and non-Ig genes.

A consensus region containing an essential SHM targeting element

downstream of the previously identified transcriptional enhancer

emerges, that is consistent with findings described in all three reports

(Fig. 3.2). The second half of the bipartite DIVAC (5′ of the

transcriptional enhancer, see Fig. 3.2) identified by the Buerstedde

group, was not detected in any of the other studies. This is likely to be

due to differences in the experimental approach, and its importance

needs to be addressed by deleting this sequences from an otherwise


DNA Deamination and the Immune System

Figure 3.2. Targeting elements in the chicken IgL locus. In this schematic representation

of a rearranged chicken IgL gene locus the exons are shown as black boxes, the

previously reported enhancer as an oval, and the CR1 retrotransposon as a grey box. The

regions within the locus that have been shown to be important for targeting of AIDmediated sequence diversification are shown as black lines below the gene. Note that the

diversification activator (DIVAC) (Blagodatski et al., 2009) and Frag. A have been

identified using reporter genes (Kim and Tian, 2009), while the 3′RR and the mutation

enhancer element (MEE) have be found based on their importance for endogenous VJ

exon mutations (Kothapalli et al., 2008; N. Kothapalli and S.D. Fugmann, unpublished

data). The consensus region corresponds to nucleotides 1147969-1148336 of chromosome

15 in the chicken genome (Gallus gallus assembly 2.1, NW_001471461).

unaltered IgL gene locus. Overall, the DT40 IgL locus model system

provides strong support for the existence of cis-acting targeting elements,

and also suggests that the SHM targeting requires sequence elements

distinct from those for enhancing transcription.

Given the size of classic cis-regulatory elements (enhancers, silencers

and insulators), being in the order of a few hundred base pairs and

containing multiple distinct transcription-factor binding sites, it is likely

that the targeting elements are of similar size and exhibit similar features.

Standard transcription-factor binding site prediction programs reveal a

large number of putative binding sites within the IgL MEE (S.D.

Fugmann, unpublished data), but the relevance of each of these sites

remains to be determined. Interestingly, E2A sites are present in all

murine Ig loci and were also found to be enriched in the vicinity of the

non-Ig genes undergoing AID-dependent mutagenesis (Liu et al., 2008).

E2A sites also act as enhancers of SHM when present in SHM reporter

transgenes in mice (Michael et al., 2003). Lastly, a recent report

suggested that NFκB, octamer and Mef2 binding sites might be

components of the targeting element in the DT40 IgL locus, as altering

these sequences led to a significant reduction in gene conversion in the

context of an artificial reporter construct (Kim and Tian, 2009).

Cis-Regulatory Elements that Target AID to Immunoglobulin Loci


However, the impact of mutating any such individual binding sites has

not been assessed in the context of the endogenous IgL gene locus thus

far. On the other hand, DT40 cells deficient in respective transcription

factors (c-Rel, p50, E2A) showed reduced frequencies of SHM/GCV

(Kim and Tian, 2009; Schoetz et al., 2006), but it is unclear at this point

whether they act directly by binding to the IgL locus, or indirectly by

controlling the transcription of genes involved in these processes.

3.4 Targeting Elements in the Murine IgH Locus

3.4.1 Targeting of CSR

CSR, in its simplest conception, is a DNA deletion that begins in a

region between the VDJ exon and the Cµ exon, and ends upstream of one

of the Cγ, Cε or Cα genes. The regions in which the deletion begins and

ends have been called “S regions” for their role in switch recombination

(Fig. 3.3). S regions are composed of simple sequences repeated in

tandem over 2−9kb, and lie about 3kb 5′ of each of the CH gene, except

for Cδ. The consequence of the 40−170kb switch deletion is to move the

VDJ exon into physical and functional association with a new CH gene,

so that the exon encoding the antigen-binding domain is now associated

with a new effector function (i.e. a new CH region). The ends of the

deleted DNA are also ligated to one another, forming a circle (von

Schwedler et al., 1990; Iwasato et al., 1990; Matsuoka et al., 1990), and

so the class switch is a quasi-reciprocal recombination event. In fact,

class switching is most often referred to as a “recombination” event,

emphasizing that it involves the ligation of two pieces of DNA.

Figure 3.3. The murine IgH locus. In this schematic representation of the 3′ end of the

murine IgH locus, the rearranged VDJ exon and the constant (C) region exons are shown

as grey boxes, the switch (S) regions as open boxes, and the enhancer (E) elements and

hypersensitive sites (HS) as black diamonds. The size and distance of important elements

is shown above the gene locus.


DNA Deamination and the Immune System

In order for CSR to occur, AID must deaminate cytosines in S regions

(and not in other DNA; Longerich et al., 2006; Chaudhuri and Alt,

2004). Next, repair/recombination enzymes must process the U in DNA

to create what is likely to be a double-stranded break (usually with short

or long single-stranded flaps; Longerich et al., 2006; Chaudhuri and Alt,

2004). It is also likely that one type of processing is to fill in the singlestranded flaps by error-prone DNA synthesis, resulting in flush doublestranded breaks (Chen et al., 2001; Chaudhuri and Alt, 2004; Longerich

et al., 2006). Two double-stranded breaks, one from the Sµ region and

one from the Sγ, Sε or Sα region must synapse. Finally, the two breaks

must be ligated together if CSR is to actually exchange CH regions. A

goal of many investigators has been to identify the cis-acting elements

that would help bring all these factors to S regions and would promote S

region synapsis. Enhancers of transcription as enhancers

of recombination

Transcriptional enhancers have been logical candidates for elements that

would also improve CSR. Enhancers have a clear role in the regulation

of V(D)J recombination (Sleckman et al., 1996). Indeed, the enhancer

found between the VDJ exon and Sµ (termed “Eµ”) has a measurable,

albeit modest, effect on CSR (Gu et al., 1993; Bottaro et al., 1998; Sakai

et al., 1999). Normally, CSR is very efficient: virtually all hybridomas

expressing IgG, both heavy chain alleles have undergone some DNA

rearrangement, and this is usually a deletion between Sµ and Sγ.

However, 25% of heavy chain alleles with a deletion of the Eµ region

fail to undergo a recombination event, and remain in the germline

configuration. CSR remains efficient for most alleles, even in the

absence of Eµ, and so additional candidates for enhancement of CSR

were sought (Gu et al., 1993; Bottaro et al., 1998).

Four DNAse I hypersensitive sites lie 3′ of the Cα gene in mice,

called HS3A, HS1,2, HS3B and HS4 in their order from most Cα

proximal to distal (Fig. 3.3; Cogne and Birshtein 2004). HS1,2 represents

two DNAse hypersensitive sites that are so close to each other that they

are always evaluated as a single regulatory element. HS3A and HS3B

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 Targeting by Ig Promoters – Are High Levels ofTranscription All There is to It?

Tải bản đầy đủ ngay(0 tr)