Limits...
Detailed structural-functional analysis of the Krüppel-like factor 16 (KLF16) transcription factor reveals novel mechanisms for silencing Sp/KLF sites involved in metabolism and endocrinology.

Daftary GS, Lomberk GA, Buttar NS, Allen TW, Grzenda A, Zhang J, Zheng Y, Mathison AJ, Gada RP, Calvo E, Iovanna JL, Billadeau DD, Prendergast FG, Urrutia R - J. Biol. Chem. (2011)

Bottom Line: We found that KLF16 selectively binds three distinct KLF-binding sites (GC, CA, and BTE boxes).Thus, this study lends insights on key biochemical mechanisms for regulating KLF sites involved in reproductive biology.These data also contribute to the new functional information that is applicable to understanding KLF16 and other highly related KLF proteins.

View Article: PubMed Central - PubMed

Affiliation: Department of Obstetrics and Gynecology, Mayo Clinic, Rochester, Minnesota 55905, USA. daftary.gaurang@mayo.edu

ABSTRACT
Krüppel-like factor (KLF) proteins have elicited significant attention due to their emerging key role in metabolic and endocrine diseases. Here, we extend this knowledge through the biochemical characterization of KLF16, unveiling novel mechanisms regulating expression of genes involved in reproductive endocrinology. We found that KLF16 selectively binds three distinct KLF-binding sites (GC, CA, and BTE boxes). KLF16 also regulated the expression of several genes essential for metabolic and endocrine processes in sex steroid-sensitive uterine cells. Mechanistically, we determined that KLF16 possesses an activation domain that couples to histone acetyltransferase-mediated pathways, as well as a repression domain that interacts with the histone deacetylase chromatin-remodeling system via all three Sin3 isoforms, suggesting a higher level of plasticity in chromatin cofactor selection. Molecular modeling combined with molecular dynamic simulations of the Sin3a-KLF16 complex revealed important insights into how this interaction occurs at an atomic resolution level, predicting that phosphorylation of Tyr-10 may modulate KLF16 function. Phosphorylation of KLF16 was confirmed by in vivo (32)P incorporation and controlled by a Y10F site-directed mutant. Inhibition of Src-type tyrosine kinase signaling as well as the nonphosphorylatable Y10F mutation disrupted KLF16-mediated gene silencing, demonstrating that its function is regulatable rather than constitutive. Subcellular localization studies revealed that signal-induced nuclear translocation and euchromatic compartmentalization constitute an additional mechanism for regulating KLF16 function. Thus, this study lends insights on key biochemical mechanisms for regulating KLF sites involved in reproductive biology. These data also contribute to the new functional information that is applicable to understanding KLF16 and other highly related KLF proteins.

Show MeSH

Related in: MedlinePlus

KLF16 DNA binding domain displayed selectivity for distinct cis-regulatory elements.a, sequence alignment of the zinc fingers of SP1 and KLF16 revealing critical conserved residues (gray). Predicted KLF16 binding to GGG, GCG, and GGG bases via zinc fingers 1–3, respectively (ZF 1, 2, 3), is shown. b, ROB assay for KLF16 in uterine cells. 32P-Labeled ROB oligonucleotides were incubated with GST-KLF16 and separated by nondenaturing gel electrophoresis. The consensus sequence derived from 45 sequenced oligonucleotides after seven rounds of ROB is shown. The font size of each nucleotide corresponds to frequency of occurrence. c, EMSA using either 32P-labeled consensus KLF16-ROB (lanes 1, 2 and 5–8), BTE (lane 4), or noncompetitor probe (lane 9) with either 1 μg of GST protein (lane 2), GST-KLF16-zinc finger (ZF) (lanes 3–9), or probe alone (lane 1). Where indicated, the following were also added: 1 μg of anti-GST (lane 6), 500 m excess of unlabeled KLF16-ROB consensus probe (cold competitor, lane 7), or 500 m excess unlabeled noncompetitor probe (lane 8). Specific complexes formed between GST-KLF16 and either labeled BTE or KLF16-ROB probe (lanes 5 and 8) are indicated by an arrow. Anti-GST disrupted the GST-KLF16-ZF/consensus probe complex (lane 6). Addition of excess unlabeled consensus probe competed for the binding (lane 7), whereas an unrelated noncompetitor probe did not (lane 8). GST-KLF16-ZF did not shift the labeled noncompetitor probe (lane 9). Lane 1 contained only labeled KLF16-ROB consensus probe and no protein, whereas lane 3 contained 1 μg of GST-KLF16-ZF fusion protein alone. d, EMSA comparing GST-KLF16-ZF (1 μg per lane; lanes 1–9) protein binding to either 32P-labeled CYP1A1 BTE (lanes 1, 4, and 7), KLF consensus element (GC box) (lanes 2, 5, and 8), and CA box (lanes 3, 6, and 9). GST-KLF16-ZF protein (lanes 1–3) specifically bound all three KLF-binding elements (lanes 1–3, arrow indicates the shift), the BTE box with highest affinity (lane 1), robust binding to the GC box (lane 2), and the CA box with least affinity (lane 3). KLF16 binding to all elements was specifically disrupted by 1 μg of anti-GST (lanes 4–6: lane 4, BTE; lane 5, GC; lane 6, CA). Binding was lost on addition of 500 m fold excess unlabeled specific cold competitor (lanes 7–9: lane 7, BTE; lane 8, GC; lane 9, CA). e, uterine cells were cotransfected with 7.5 μg of FLAG-KLF16 or corresponding empty vector and either a 6×-BTE-, 6×-GC-, or 6×-CA-luciferase reporter vector (2.5 μg). Luciferase levels normalized to total protein levels and binding showed significant repression compared with empty vector, when KLF16 was cotransfected with either the 6×BTE (*, p = 0.001) or 6×GC-luciferase reporter (**, p < 0.001). There was no repression of 6×-CA-luciferase (f). HEC1A uterine cells were cotransfected for 48 h with either 7.5 μg of FLAG-KLF16 or corresponding parent pCMV/FLAG vector and 2.5 μg of luciferase reporter containing 6×-tandem BTEs. Luciferase activity normalized to protein concentrations show that compared with EV, KLF16 decreased luciferase expression by 63% (*, p = 0.003).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3293586&req=5

Figure 1: KLF16 DNA binding domain displayed selectivity for distinct cis-regulatory elements.a, sequence alignment of the zinc fingers of SP1 and KLF16 revealing critical conserved residues (gray). Predicted KLF16 binding to GGG, GCG, and GGG bases via zinc fingers 1–3, respectively (ZF 1, 2, 3), is shown. b, ROB assay for KLF16 in uterine cells. 32P-Labeled ROB oligonucleotides were incubated with GST-KLF16 and separated by nondenaturing gel electrophoresis. The consensus sequence derived from 45 sequenced oligonucleotides after seven rounds of ROB is shown. The font size of each nucleotide corresponds to frequency of occurrence. c, EMSA using either 32P-labeled consensus KLF16-ROB (lanes 1, 2 and 5–8), BTE (lane 4), or noncompetitor probe (lane 9) with either 1 μg of GST protein (lane 2), GST-KLF16-zinc finger (ZF) (lanes 3–9), or probe alone (lane 1). Where indicated, the following were also added: 1 μg of anti-GST (lane 6), 500 m excess of unlabeled KLF16-ROB consensus probe (cold competitor, lane 7), or 500 m excess unlabeled noncompetitor probe (lane 8). Specific complexes formed between GST-KLF16 and either labeled BTE or KLF16-ROB probe (lanes 5 and 8) are indicated by an arrow. Anti-GST disrupted the GST-KLF16-ZF/consensus probe complex (lane 6). Addition of excess unlabeled consensus probe competed for the binding (lane 7), whereas an unrelated noncompetitor probe did not (lane 8). GST-KLF16-ZF did not shift the labeled noncompetitor probe (lane 9). Lane 1 contained only labeled KLF16-ROB consensus probe and no protein, whereas lane 3 contained 1 μg of GST-KLF16-ZF fusion protein alone. d, EMSA comparing GST-KLF16-ZF (1 μg per lane; lanes 1–9) protein binding to either 32P-labeled CYP1A1 BTE (lanes 1, 4, and 7), KLF consensus element (GC box) (lanes 2, 5, and 8), and CA box (lanes 3, 6, and 9). GST-KLF16-ZF protein (lanes 1–3) specifically bound all three KLF-binding elements (lanes 1–3, arrow indicates the shift), the BTE box with highest affinity (lane 1), robust binding to the GC box (lane 2), and the CA box with least affinity (lane 3). KLF16 binding to all elements was specifically disrupted by 1 μg of anti-GST (lanes 4–6: lane 4, BTE; lane 5, GC; lane 6, CA). Binding was lost on addition of 500 m fold excess unlabeled specific cold competitor (lanes 7–9: lane 7, BTE; lane 8, GC; lane 9, CA). e, uterine cells were cotransfected with 7.5 μg of FLAG-KLF16 or corresponding empty vector and either a 6×-BTE-, 6×-GC-, or 6×-CA-luciferase reporter vector (2.5 μg). Luciferase levels normalized to total protein levels and binding showed significant repression compared with empty vector, when KLF16 was cotransfected with either the 6×BTE (*, p = 0.001) or 6×GC-luciferase reporter (**, p < 0.001). There was no repression of 6×-CA-luciferase (f). HEC1A uterine cells were cotransfected for 48 h with either 7.5 μg of FLAG-KLF16 or corresponding parent pCMV/FLAG vector and 2.5 μg of luciferase reporter containing 6×-tandem BTEs. Luciferase activity normalized to protein concentrations show that compared with EV, KLF16 decreased luciferase expression by 63% (*, p = 0.003).

Mentions: Our studies began by analyzing the DNA binding functions of KLF16, which are key for better understanding this protein and predicting candidate gene targets, as well as the potential competition among KLF family members that may share KLF16 DNA binding activity. Generally, KLF proteins target promoters differentially via three well characterized GC-rich elements as follows: the BTE (GAGGCGTGGCCAAC), GC box (CGGGGCGGGGC), and CA box (CACCC) (11). Interestingly, studies from DNA-bound zinc finger peptides permit prediction of putative DNA sequences that may be recognized by novel zinc finger proteins such as the KLF proteins (11, 23). The amino acid residues within the first (KHA), second (RER), and third (RHK) zinc fingers of KLF16 are identical to corresponding regions within SP1 (Fig. 1a) that bind the sequences GGG (ZF1), GCG (ZF2), and GGG (ZF3), respectively (23). Thus, a priori prediction suggests that KLF16 prefers GC-rich cis-regulatory elements over the CA box KLF sequence. To test this prediction, we performed ROB assays with a library of DNA sequences consisting of 12-bp random cores flanked bilaterally by 16 bp of known sequence, the results of which were aligned to derive a consensus sequence (Fig. 1b). KLF16 binding to this consensus oligonucleotide (GGGGGGGGGCGG) was confirmed by EMSA. Additionally, KLF16 binding specificity was validated by both supershift assay showing complex disruption with anti-GST antibodies and specific site-based competition with cold probes (Fig. 1c). Together, these experiments demonstrated that KLF16 self-selected a GC-rich sequence in vitro that was similar to, yet distinct from, previously described KLF-binding sites. Characterization of this specific DNA element was critical for subsequent identification of candidate KLF16 gene targets by genome-wide analyses of cis-regulatory sequences. Additionally, we comparatively analyzed the binding of KLF16 to the consensus probe with other KLF cis-regulatory sequences, namely the GC, CA. and BTE boxes (Fig. 1, c and d) (1, 2). Densitometric analysis of EMSA data revealed that KLF16 revealed a 5-fold greater preference for the BTE probe compared with the consensus ROB probe (Fig. 1c). In addition, a 17- and 10-fold greater preference for the GC box (very similar to the consensus ROB sequence) and the BTE, respectively, compared with the CA box was observed (Fig. 1d). KLF16 binding to all elements was specific as confirmed by supershift and cold probe competition (Fig. 1, c and d). Thus, the combination of an unbiased approach (ROB), together with three candidate-based studies using known KLF cis-regulatory domains, provided the best comparative information available for any KLF protein, as well as confirmed the a priori prediction (23). KLF16 recognized three different GC-rich sequences with varying affinity, positioning this protein as a candidate to regulate similar sites in gene promoters.


Detailed structural-functional analysis of the Krüppel-like factor 16 (KLF16) transcription factor reveals novel mechanisms for silencing Sp/KLF sites involved in metabolism and endocrinology.

Daftary GS, Lomberk GA, Buttar NS, Allen TW, Grzenda A, Zhang J, Zheng Y, Mathison AJ, Gada RP, Calvo E, Iovanna JL, Billadeau DD, Prendergast FG, Urrutia R - J. Biol. Chem. (2011)

KLF16 DNA binding domain displayed selectivity for distinct cis-regulatory elements.a, sequence alignment of the zinc fingers of SP1 and KLF16 revealing critical conserved residues (gray). Predicted KLF16 binding to GGG, GCG, and GGG bases via zinc fingers 1–3, respectively (ZF 1, 2, 3), is shown. b, ROB assay for KLF16 in uterine cells. 32P-Labeled ROB oligonucleotides were incubated with GST-KLF16 and separated by nondenaturing gel electrophoresis. The consensus sequence derived from 45 sequenced oligonucleotides after seven rounds of ROB is shown. The font size of each nucleotide corresponds to frequency of occurrence. c, EMSA using either 32P-labeled consensus KLF16-ROB (lanes 1, 2 and 5–8), BTE (lane 4), or noncompetitor probe (lane 9) with either 1 μg of GST protein (lane 2), GST-KLF16-zinc finger (ZF) (lanes 3–9), or probe alone (lane 1). Where indicated, the following were also added: 1 μg of anti-GST (lane 6), 500 m excess of unlabeled KLF16-ROB consensus probe (cold competitor, lane 7), or 500 m excess unlabeled noncompetitor probe (lane 8). Specific complexes formed between GST-KLF16 and either labeled BTE or KLF16-ROB probe (lanes 5 and 8) are indicated by an arrow. Anti-GST disrupted the GST-KLF16-ZF/consensus probe complex (lane 6). Addition of excess unlabeled consensus probe competed for the binding (lane 7), whereas an unrelated noncompetitor probe did not (lane 8). GST-KLF16-ZF did not shift the labeled noncompetitor probe (lane 9). Lane 1 contained only labeled KLF16-ROB consensus probe and no protein, whereas lane 3 contained 1 μg of GST-KLF16-ZF fusion protein alone. d, EMSA comparing GST-KLF16-ZF (1 μg per lane; lanes 1–9) protein binding to either 32P-labeled CYP1A1 BTE (lanes 1, 4, and 7), KLF consensus element (GC box) (lanes 2, 5, and 8), and CA box (lanes 3, 6, and 9). GST-KLF16-ZF protein (lanes 1–3) specifically bound all three KLF-binding elements (lanes 1–3, arrow indicates the shift), the BTE box with highest affinity (lane 1), robust binding to the GC box (lane 2), and the CA box with least affinity (lane 3). KLF16 binding to all elements was specifically disrupted by 1 μg of anti-GST (lanes 4–6: lane 4, BTE; lane 5, GC; lane 6, CA). Binding was lost on addition of 500 m fold excess unlabeled specific cold competitor (lanes 7–9: lane 7, BTE; lane 8, GC; lane 9, CA). e, uterine cells were cotransfected with 7.5 μg of FLAG-KLF16 or corresponding empty vector and either a 6×-BTE-, 6×-GC-, or 6×-CA-luciferase reporter vector (2.5 μg). Luciferase levels normalized to total protein levels and binding showed significant repression compared with empty vector, when KLF16 was cotransfected with either the 6×BTE (*, p = 0.001) or 6×GC-luciferase reporter (**, p < 0.001). There was no repression of 6×-CA-luciferase (f). HEC1A uterine cells were cotransfected for 48 h with either 7.5 μg of FLAG-KLF16 or corresponding parent pCMV/FLAG vector and 2.5 μg of luciferase reporter containing 6×-tandem BTEs. Luciferase activity normalized to protein concentrations show that compared with EV, KLF16 decreased luciferase expression by 63% (*, p = 0.003).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3293586&req=5

Figure 1: KLF16 DNA binding domain displayed selectivity for distinct cis-regulatory elements.a, sequence alignment of the zinc fingers of SP1 and KLF16 revealing critical conserved residues (gray). Predicted KLF16 binding to GGG, GCG, and GGG bases via zinc fingers 1–3, respectively (ZF 1, 2, 3), is shown. b, ROB assay for KLF16 in uterine cells. 32P-Labeled ROB oligonucleotides were incubated with GST-KLF16 and separated by nondenaturing gel electrophoresis. The consensus sequence derived from 45 sequenced oligonucleotides after seven rounds of ROB is shown. The font size of each nucleotide corresponds to frequency of occurrence. c, EMSA using either 32P-labeled consensus KLF16-ROB (lanes 1, 2 and 5–8), BTE (lane 4), or noncompetitor probe (lane 9) with either 1 μg of GST protein (lane 2), GST-KLF16-zinc finger (ZF) (lanes 3–9), or probe alone (lane 1). Where indicated, the following were also added: 1 μg of anti-GST (lane 6), 500 m excess of unlabeled KLF16-ROB consensus probe (cold competitor, lane 7), or 500 m excess unlabeled noncompetitor probe (lane 8). Specific complexes formed between GST-KLF16 and either labeled BTE or KLF16-ROB probe (lanes 5 and 8) are indicated by an arrow. Anti-GST disrupted the GST-KLF16-ZF/consensus probe complex (lane 6). Addition of excess unlabeled consensus probe competed for the binding (lane 7), whereas an unrelated noncompetitor probe did not (lane 8). GST-KLF16-ZF did not shift the labeled noncompetitor probe (lane 9). Lane 1 contained only labeled KLF16-ROB consensus probe and no protein, whereas lane 3 contained 1 μg of GST-KLF16-ZF fusion protein alone. d, EMSA comparing GST-KLF16-ZF (1 μg per lane; lanes 1–9) protein binding to either 32P-labeled CYP1A1 BTE (lanes 1, 4, and 7), KLF consensus element (GC box) (lanes 2, 5, and 8), and CA box (lanes 3, 6, and 9). GST-KLF16-ZF protein (lanes 1–3) specifically bound all three KLF-binding elements (lanes 1–3, arrow indicates the shift), the BTE box with highest affinity (lane 1), robust binding to the GC box (lane 2), and the CA box with least affinity (lane 3). KLF16 binding to all elements was specifically disrupted by 1 μg of anti-GST (lanes 4–6: lane 4, BTE; lane 5, GC; lane 6, CA). Binding was lost on addition of 500 m fold excess unlabeled specific cold competitor (lanes 7–9: lane 7, BTE; lane 8, GC; lane 9, CA). e, uterine cells were cotransfected with 7.5 μg of FLAG-KLF16 or corresponding empty vector and either a 6×-BTE-, 6×-GC-, or 6×-CA-luciferase reporter vector (2.5 μg). Luciferase levels normalized to total protein levels and binding showed significant repression compared with empty vector, when KLF16 was cotransfected with either the 6×BTE (*, p = 0.001) or 6×GC-luciferase reporter (**, p < 0.001). There was no repression of 6×-CA-luciferase (f). HEC1A uterine cells were cotransfected for 48 h with either 7.5 μg of FLAG-KLF16 or corresponding parent pCMV/FLAG vector and 2.5 μg of luciferase reporter containing 6×-tandem BTEs. Luciferase activity normalized to protein concentrations show that compared with EV, KLF16 decreased luciferase expression by 63% (*, p = 0.003).
Mentions: Our studies began by analyzing the DNA binding functions of KLF16, which are key for better understanding this protein and predicting candidate gene targets, as well as the potential competition among KLF family members that may share KLF16 DNA binding activity. Generally, KLF proteins target promoters differentially via three well characterized GC-rich elements as follows: the BTE (GAGGCGTGGCCAAC), GC box (CGGGGCGGGGC), and CA box (CACCC) (11). Interestingly, studies from DNA-bound zinc finger peptides permit prediction of putative DNA sequences that may be recognized by novel zinc finger proteins such as the KLF proteins (11, 23). The amino acid residues within the first (KHA), second (RER), and third (RHK) zinc fingers of KLF16 are identical to corresponding regions within SP1 (Fig. 1a) that bind the sequences GGG (ZF1), GCG (ZF2), and GGG (ZF3), respectively (23). Thus, a priori prediction suggests that KLF16 prefers GC-rich cis-regulatory elements over the CA box KLF sequence. To test this prediction, we performed ROB assays with a library of DNA sequences consisting of 12-bp random cores flanked bilaterally by 16 bp of known sequence, the results of which were aligned to derive a consensus sequence (Fig. 1b). KLF16 binding to this consensus oligonucleotide (GGGGGGGGGCGG) was confirmed by EMSA. Additionally, KLF16 binding specificity was validated by both supershift assay showing complex disruption with anti-GST antibodies and specific site-based competition with cold probes (Fig. 1c). Together, these experiments demonstrated that KLF16 self-selected a GC-rich sequence in vitro that was similar to, yet distinct from, previously described KLF-binding sites. Characterization of this specific DNA element was critical for subsequent identification of candidate KLF16 gene targets by genome-wide analyses of cis-regulatory sequences. Additionally, we comparatively analyzed the binding of KLF16 to the consensus probe with other KLF cis-regulatory sequences, namely the GC, CA. and BTE boxes (Fig. 1, c and d) (1, 2). Densitometric analysis of EMSA data revealed that KLF16 revealed a 5-fold greater preference for the BTE probe compared with the consensus ROB probe (Fig. 1c). In addition, a 17- and 10-fold greater preference for the GC box (very similar to the consensus ROB sequence) and the BTE, respectively, compared with the CA box was observed (Fig. 1d). KLF16 binding to all elements was specific as confirmed by supershift and cold probe competition (Fig. 1, c and d). Thus, the combination of an unbiased approach (ROB), together with three candidate-based studies using known KLF cis-regulatory domains, provided the best comparative information available for any KLF protein, as well as confirmed the a priori prediction (23). KLF16 recognized three different GC-rich sequences with varying affinity, positioning this protein as a candidate to regulate similar sites in gene promoters.

Bottom Line: We found that KLF16 selectively binds three distinct KLF-binding sites (GC, CA, and BTE boxes).Thus, this study lends insights on key biochemical mechanisms for regulating KLF sites involved in reproductive biology.These data also contribute to the new functional information that is applicable to understanding KLF16 and other highly related KLF proteins.

View Article: PubMed Central - PubMed

Affiliation: Department of Obstetrics and Gynecology, Mayo Clinic, Rochester, Minnesota 55905, USA. daftary.gaurang@mayo.edu

ABSTRACT
Krüppel-like factor (KLF) proteins have elicited significant attention due to their emerging key role in metabolic and endocrine diseases. Here, we extend this knowledge through the biochemical characterization of KLF16, unveiling novel mechanisms regulating expression of genes involved in reproductive endocrinology. We found that KLF16 selectively binds three distinct KLF-binding sites (GC, CA, and BTE boxes). KLF16 also regulated the expression of several genes essential for metabolic and endocrine processes in sex steroid-sensitive uterine cells. Mechanistically, we determined that KLF16 possesses an activation domain that couples to histone acetyltransferase-mediated pathways, as well as a repression domain that interacts with the histone deacetylase chromatin-remodeling system via all three Sin3 isoforms, suggesting a higher level of plasticity in chromatin cofactor selection. Molecular modeling combined with molecular dynamic simulations of the Sin3a-KLF16 complex revealed important insights into how this interaction occurs at an atomic resolution level, predicting that phosphorylation of Tyr-10 may modulate KLF16 function. Phosphorylation of KLF16 was confirmed by in vivo (32)P incorporation and controlled by a Y10F site-directed mutant. Inhibition of Src-type tyrosine kinase signaling as well as the nonphosphorylatable Y10F mutation disrupted KLF16-mediated gene silencing, demonstrating that its function is regulatable rather than constitutive. Subcellular localization studies revealed that signal-induced nuclear translocation and euchromatic compartmentalization constitute an additional mechanism for regulating KLF16 function. Thus, this study lends insights on key biochemical mechanisms for regulating KLF sites involved in reproductive biology. These data also contribute to the new functional information that is applicable to understanding KLF16 and other highly related KLF proteins.

Show MeSH
Related in: MedlinePlus