In vivo analysis of interactions between trans-acting factors and their target genes

The investigations presented in this thesis use various in vivo techniques to understand how trans-acting factors control gene expression. The first part addresses the transcriptional regulation of muscle creatine kinase (MCK). MCK expression is activated during the course of development and is found only in differentiated muscle. Several in vivo footprints are observed at the enhancer of this gene, but all of these interactions are limited to cell types that express MCK. This is interesting because two of the footprints appear to represent muscle specific use of general transcription factors, while the other two correspond to sites that can bind the myogenic regulator, MyoD1, in vitro. MyoD1 and these general factors are present in myoblasts, but can bind to the enhancer only in myocytes. This suggests that either the factors themselves are post-translationally modified (phosphorylation or protein:protein interactions), or the accessibility of the enhancer to the factors is limited (changes in chromatin structure). The in vivo footprinting study of MCK was performed with a new ligation mediated, single-sided PCR (polymerase chain reaction) technique that I have developed.

The second half of the thesis concerns the regulation of mouse metallothionein (MT). Metallothioneins are a family of highly conserved housekeeping genes whose expression can be induced by heavy metals, steroids, and other stresses. By adapting a primer extension method of genomic sequencing to in vivo footprinting, I've observed both metal inducible and noninducible interactions at the promoter of MT-I. From these results I've been able to limit the possible mechanisms by which metal responsive trans-acting factors induce transcription. These interpretations correlate with a second line of experiments involving the stable titration of positive acting factors necessary for induction of MT. I've amplified the promoter of MT to 10^2-10^3 copies per cell by fusing the 5' and 3' ends of the MT gene to the coding region of DHFR and selecting cells for methotrexate resistance. In these cells, there is a metal-specific titration effect, and although it acts at the level of transcription, it appears to be independent of direct DNA binding factors.

Análise de polimorfismos e de expressão do gene p16INK4a em neoplasias cervicais

O câncer de colo do útero é o segundo carcinoma mais frequente em mulheres no mundo e um dos cânceres femininos mais incidentes no Brasil. Em lesões pré-malignas e malignas do colo uterino, a proteína p16INK4a, que participa do controle do ciclo celular, apresenta um aumento considerável de sua expressão, devido possivelmente à presença de oncoproteínas do papilomavírus humano (HPV). Dois polimorfismos no gene p16INK4a, p16 500C>G e p16 540C>T, estão localizados na região 3 não traduzida (3UTR), que está envolvida na regulação pós-transcricional da expressão gênica. O objetivo deste estudo foi avaliar possíveis associações entre os polimorfismos p16 500C>G e p16 540C>T e o desenvolvimento de neoplasias cervicais e/ou a severidade das lesões, considerando os níveis de expressão da proteína p16INK4a nas lesões cervicais e certos fatores de risco clássicos para o câncer cervical, incluindo a infecção pelo HPV. Para isso, foram selecionadas 567 mulheres residentes no Rio de Janeiro, 319 com citologia cervical alterada (grupo de casos) e 248 sem história prévia de alteração citológica do colo uterino (grupo de comparação). Amostras de sangue periférico de todas as participantes foram utilizadas na análise molecular dos polimorfismos p16 500C>G e p16 540C>T através da técnica de PCR-RFLP (reação em cadeia da polimerase - polimorfismo de comprimento de fragmento de restrição), usando as enzimas de restrição MspI e HaeIII, respectivamente. A expressão da proteína p16INK4a em 137 biópsias de mulheres pertencentes ao grupo de casos foi avaliada por imunohistoquímica. A detecção de DNA do HPV em células cervicais foi feita em todas as amostras do grupo de comparação e em 194 amostras do grupo de casos pela técnica de PCR, usando dois pares de oligonucleotídeos, MY09/MY11 e GP05+/GP06+. Os dois grupos de estudo se encontram em equilíbrio de Hardy-Weinberg. As distribuições genotípicas para p16 500C>G e p16 540C>T e as distribuições de combinações haplotípicas nos dois grupos não apresentaram diferenças significativas. A análise do subgrupo HSIL+câncer (casos com lesão intraepitelial de alto grau ou carcinoma invasivo) em comparação com o subgrupo LSIL (casos com lesão intraepitelial de baixo grau) revelou diferença significativa entre as distribuições das combinações haplotípicas (p = 0,036) e diferenças marginais entre as distribuições genotípicas para p16 500C>G (p = 0,071) e p16 540C>T (p = 0,051). O alelo p16 540G, em heterozigose ou homozigose (OR = 1,91, IC 95% = 1,08-3,37), e a combinação haplotípica p16 500C-540C 500G-540C (OR = 2,34, IC 95% = 1,202-4,555) mostraram-se associados com a severidade da lesões cervicais. Já o genótipo p16 540T/T (OR = 0,25, IC 95% = 0,08-0,79), e a combinação haplotípica p16 500C-540T 500C-540T (OR = 0,27, IC 95% = 0,088-0,827) exibiram papel protetor contra o desenvolvimento de lesões mais severas. As análises de interação entre os polimorfismos de p16INK4a e a expressão de p16 ou a infecção pelo HPV foram comprometidas pelo número reduzido de amostras analisadas. Não se observou qualquer interação entre os polimorfismos estudados e os fatores de risco clássicos para o câncer de colo uterino. Nossos resultados apontam para a importância dos polimorfismos do gene p16INK4a como marcadores de severidade da neoplasia cervical.

Estudo da cromatina nos sítios de fixação à matriz nuclear no domínio do gene TP53 e das modificações epigenéticas e no modelo de progressão tumoral mamária 21T

Dentre os diversos tipos de câncer agressivos, o câncer de mama é o mais comum em mulheres. Mutações hereditárias e adquiridas, assim como alterações epigenéticas atuam em sinergia na carcinogênese mamária e na progressão tumoral. A proteína P53 é uma supressora de tumor e possui uma atuação fundamental na integridade genômica. Apesar do vasto conhecimento sobre o controle da P53 a nível de proteína, ainda pouco se sabe sobre o controle transcricional do gene TP53. A série 21T, uma série de 4 linhagens celulares originadas da mama da mesma paciente, representando diferentes estágios de progressão tumoral mamária, é um eficiente modelo para investigação das alterações epigenéticas e suas influências na expressão gênica ao longo da progressão do câncer de mama. Nós analisamos a organização do domínio do gene TP53 através da técnica de arranjo de DNA, em diversas linhagens celulares de câncer de mama e linhagens controle, e realizamos uma tentativa de caracterizar estes elementos de DNA nas linhagens controle não-tumorais HB2 e MCF10A e nas tumorais MCF-7, MDA-MB-231, T47D, através dos marcadores epigenéticos de eucromatina, H4Ac, e heterocromatina, H3K9me3. Ainda analisamos a ligação de proteínas à região associada à matriz nuclear (MAR), denominada MAR 2, e a possível ligação da proteína ligante à matriz nuclear (MARBP), PARP-1, através de ensaios de gel shift (EMSA). Detectamos que na linhagem controle epitelial mamária, HB2, o gene TP53 está posicionado num domínio de DNA relativamente pequeno, aproximadamente 50 kb, delimitado por dois sítios de fixação à matriz nuclear. Interessantemente, esta estrutura de domínio se apresentou radicalmente diferente nas linhagens de câncer de mama estudadas, MCF7, T47D, MDA-MB-231 e BT474, nos quais o tamanho do domínio estudado estava aumentado e a transcrição do TP53 diminuída. Os enriquecimentos com os marcadores epigenéticos de cromatina H4Ac e H3K9me3 estão diferentemente distribuídos nas MARs nas linhagens celulares. Surpreendentemente, a MAR 2 apresentou uma ligação altamente específica, o que poderia representar a atuação de fatores transcricionais envolvidos na organização da cromatina. Através de programas de bioinformática, detectamos putativos sítios para interessantes fatores de transcrição, tais como o c/EBP-beta e c-myb, que poderiam atuar em cis regulando a expressão do gene TP53 e outros flanqueadores. Nós propusemos um modelo para a organização da cromatina na região de domínio do gene TP53 com os genes flanqueadores. Através da série 21T, detectamos uma hipometilação global genômica, nas células cancerosas 21NT e 21MT1. Uma importante diminuição da expressão global do marcador H4Ac nas células metastáticas 21MT1, foi detectada em relação às outras linhagens. Os níveis de RNAm das principais enzimas relacionadas as modificações epigenéticas são consistentes com as observadas hipometilação genômica e hipoacetilação. Através de microscopia confocal, verificamos que o marcador H4Ac está localizado, na maior parte na periferia e o marcador H3K9me3, pericêntrico nos núcleos tumorais. Por fim, verificamos que o promotor P1 do gene TP53 apresenta um estado de cromatina aberta, e a expressão do gene TP53 é similar em todas as células da série 21T.

MicroRNAs in breast cancer progression and DNA damage response

Os tumores de mama são caracterizados pela sua alta heterogeneidade. O câncer de mama é uma doença complexa, que possui o seu desenvolvimento fortemente influenciado por fatores ambientais, combinada a uma progressiva acumulação de mutações genéticas e desregulação epigenética de vias críticas. Alterações nos padrões de expressão gênica podem ser resultado de uma desregulação no controle de eventos epigenéticos, assim como, na regulação pós-transcricional pelo mecanismo de RNA de interferência endógeno via microRNA (miRNA). Estes eventos são capazes de levar à iniciação, à promoção e à manutenção da carcinogênese, como também ter implicações no desenvolvimento da resistência à terapia Os miRNAs formam uma classe de RNAs não codificantes, que durante os últimos anos surgiram como um dos principais reguladores da expressão gênica, através da sua capacidade de regular negativamente a atividade de RNAs mensageiros (RNAms) portadores de uma seqüencia parcialmente complementar. A importância da regulação mediada por miRNAs foi observada pela capacidade destas moléculas em regular uma vasta gama de processos biológicos incluindo a proliferação celular, diferenciação e a apoptose. Para avaliar a expressão de miRNAs durante a progressão tumoral, utilizamos como modelo experimental a série 21T que compreende 5 linhagens celulares originárias da mesma paciente diagnosticada com um tumor primário de mama do tipo ErbB2 e uma posterior metástase pulmonar. Essa série é composta pela linhagem obtida a partir do tecido normal 16N, pelas linhagens correspondentes ao carcinoma primário 21PT e 21NT e pelas linhagens obtidas um ano após o diagnóstico inicial, a partir da efusão pleural no sítio metastatico 21MT1 e 21MT2. O miRNAoma da série 21T revelou uma redução significativa nos níveis de miR-205 e nos níveis da proteina e-caderina e um enriquecimento do fator pró-metastático ZEB-1 nas células 21MT. Considerando a importância dos miRNAs na regulação da apoptose, e que a irradiação em diferentes espectros é comumente usada em procedimentos de diagnóstico como mamografia e na radioterapia, avaliamos a expressão de miRNAs após irradiação de alta e baixa energia e do tratamento doxorrubicina. Para os ensaios foram utilizados as linhagens não tumorais MCF-10A e HB-2 e as linhagens de carcinoma da mama MCF-7 e T-47D. Observou-se que raios-X de baixa energia são capazes de promover quebras na molécula do DNA e apoptose assim como, alterar sensivelmente miRNAs envolvidos nessas vias como o let-7a, miR-34a e miR-29b. No que diz respeito à resposta a danos genotóxicos, uma regulação positiva sobre a expressão de miR-29b, o qual em condições normais é regulado negativamente foi observada uma regulação positiva sobre miR-29b expressão após todos os tratamentos em células tumorais. Nossos resultados indicam que miR-29b é um possível biomarcador de estresse genotóxico e que miR-205 pode participar no potencial metastático das células 21T.

Global analysis of mRNA half-lives and de novo transcription in a dinoflagellate, Karenia brevis

Dinoflagellates possess many physiological processes that appear to be under post-transcriptional control. However, the extent to which their genes are regulated post-transcriptionally remains unresolved. To gain insight into the roles of differential mRNA stability and de novo transcription in dinoflagellates, we biosynthetically labeled RNA with 4-thiouracil to isolate newly transcribed and pre-existing RNA pools in Karenia brevis. These isolated fractions were then used for analysis of global mRNA stability and de novo transcription by hybridization to a K. brevis microarray. Global K. brevis mRNA half-lives were calculated from the ratio of newly transcribed to pre-existing RNA for 7086 array features using the online software HALO (Half-life Organizer). Overall, mRNA half-lives were substantially longer than reported in other organisms studied at the global level, ranging from 42 minutes to greater than 144 h, with a median of 33 hours. Consistent with well-documented trends observed in other organisms, housekeeping processes, including energy metabolism and transport, were significantly enriched in the most highly stable messages. Shorter-lived transcripts included a higher proportion of transcriptional regulation, stress response, and other response/regulatory processes. One such family of proteins involved in post-transcriptional regulation in chloroplasts and mitochondria, the pentatricopeptide repeat (PPR) proteins, had dramatically shorter half-lives when compared to the arrayed transcriptome. As transcript abundances for PPR proteins were previously observed to rapidly increase in response to nutrient addition, we queried the newly synthesized RNA pools at 1 and 4 h following nitrate addition to N-depleted cultures. Transcriptome-wide there was little evidence of increases in the rate of de novo transcription during the first 4 h, relative to that in N-depleted cells, and no evidence for increased PPR protein transcription. These results lend support to the growing consensus of post-transcriptional control of gene expression in dinoflagellates.

Sox2基因3'非翻译区保守元件对基因表达的调控作用

转录因子Sox2是脊椎动物早期发育中最早表达的神经系统特异性基因之一,同时在干细胞的维持中也起着关键作用.通过生物信息学分析,作者发现在脊椎动物Sox2 mRNA 3'非翻译区中存在4段非常保守的富含AU的区域.将这些片段按照不同的组合克隆到GFP和荧光素酶两种报告基因载体中,在非洲爪蟾胚胎和培养细胞中检测了这些片段对报告基因表达的影响.结果显示,Sox2的3'UTR可影响报告基因的表达水平,特别是其中的保守片段2可显著提高报告基因的表达水平,表明Sox2 3'非翻译区有可能参与Sox2表达的转录后调控.

Differential expression of the Brunol/CELF family genes during Xenopus laevis early development

The BRUNOL/CELF family of RNA-binding proteins plays important roles in post-transcriptional regulation and has been implicated in several developmental processes. In this study, we describe the cloning and expression patterns of five Brunol genes in Xenopus laevis. Among them, only Brunol2 is maternally expressed and the zygotic expression of the other four Brunol genes starts at different developmental stages. During Xenopus development, Brunol1, 4-5 are exclusively expressed in the nervous system including domains in the brain, spinal cord, optic and otic vesicles. Brunol2 and 3 are expressed in both the somatic mesoderm and the nervous system. Brunol2 is also extensively expressed in the lens. In transfected Hela cells, BRUNOL1, 2 and 3 proteins are localized in both the cytoplasm and the nucleus, while BRUNOL4 and 5 are only present in the cytoplasm, indicating their different functions.

Deep RNA sequencing at single base-pair resolution reveals high complexity of the rice transcriptome

Understanding the dynamics of eukaryotic transcriptome is essential for studying the complexity of transcriptional regulation and its impact on phenotype. However, comprehensive studies of transcriptomes at single base resolution are rare, even for modern organisms, and lacking for rice. Here, we present the first transcriptome atlas for eight organs of cultivated rice. Using high-throughput paired-end RNA-seq, we unambiguously detected transcripts expressing at an extremely low level, as well as a substantial number of novel transcripts, exons, and untranslated regions. An analysis of alternative splicing in the rice transcriptome revealed that alternative cis-splicing occurred in similar to 33% of all rice genes. This is far more than previously reported. In addition, we also identified 234 putative chimeric transcripts that seem to be produced by trans-splicing, indicating that transcript fusion events are more common than expected. In-depth analysis revealed a multitude of fusion transcripts that might be by-products of alternative splicing. Validation and chimeric transcript structural analysis provided evidence that some of these transcripts are likely to be functional in the cell. Taken together, our data provide extensive evidence that transcriptional regulation in rice is vastly more complex than previously believed.

Transcription rates of yeast genes are influenced by the distribution of introns

A comparative analysis on the intron sequence oligonucleotide usages in two sets of yeast genes with higher and lower transcription frequencies, respectively, has shown that the intron sequence structures of the two sets of genes are different. There are more potential binding sites for transcription factors in the introns of the genes with high transcription frequencies. So it is speculated that introns regulate the transcription of genes. But more evidences are needed to favor this speculation. The detailed comparative analyses on the distribution ( length and position) of introns and exons in the two sets of gene sequences also show that there is an obvious boundary between the lengths of the two sets of introns. There is no boundary between the lengths of the two sets of exons, although the means of their lengths are of discrepancy. The situation of the gene lengths ( length of intron and exon) is similar to exon lengths. As far as the relative position, the introns in two sets of genes all have a bias toward the 5' ends of genes. But as the actual position is considered, more introns in high transcription genes have a tendency to be located toward the 5' ends of genes, some even located at 5'-UTR. These results suggest that the gene transcription rates are related to the length of intron, but not to the lengths of exons and genes sequences. The positions of introns may also influence the transcription rates. The transcriptional regulation of introns may be correlative with the transcriptional regulation of the upstream of genes, or be its continuous action.

非洲爪蟾早期发育中的转录后调控机制及眼睛发育相关基因的表达研究

近来的研究表明，转录后调控对于调节脊椎动物发育过程中的细胞分化，细胞分裂及基因区域特异性表达都具有重要作用。转录后调控包括对mRNA稳定性、翻译效率、细胞内定位及poly(A)水平的调控等。Sox2基因是脊椎动物早期发育中最早表达的神经系统特异性基因之一，是脊椎动物早期神经系统发育的重要调节因子。通过生物信息学分析，我们发现，在脊椎动物Sox2 mRNA 3’非翻译区中存在4段非常保守的富含AU的区域，通过报告基因分析等手段研究发现Sox2 3’非翻译区中的部分元件可显著提高报告基因表达，提示我们Sox2的表达可能受到转录后调控。 我们通过对爪蟾Xfhl3基因序列分析时发现其3’非翻译区存在一段保守的只在两栖类具有的序列，我们克隆并检测了该基因的表达图式，并采用报告基因分析等手段研究了Xfhl3基因 3’非翻译区对报告基因表达的影响。结果发现其3’UTR可抑制报告基因表达水平。由于Xfhl3基因3’UTR中这段序列只在爪蟾基因中高度保守，而在在进化过程中两栖动物最独特的便是变态现象，这提示我们去探索这段爪蟾特有的保守序列是否与两栖类变态发育密切相关。由于甲状腺激素在两栖类的变态中的重要作用，因此我们设想Xfhl3基因的3’UTR中的保守序列可能与甲状腺激素相互作用共同调节爪蟾的变态过程。我们的初步结果表明，在爪蟾胚胎中，甲状腺激素对于正常报告基因表达没有明显的作用，但是在插入Xfhl3基因3’UTR中保守序列后，甲状腺激素处理可显著提高报告基因的表达，表明甲状腺素可能直接或间接通过与该段保守序列参与基因的表达调控。 脊椎动物的眼是一个功能非常特殊的器官，受到复杂的调控网络的调节，众多对神经发生重要的基因在眼中表达并参与了这一调节过程。我们克隆了非洲爪蟾的Sox1基因并研究了它在非洲爪蟾早期发育过程中的时空表达图式，比较了Sox1-3基因在发育的脑和眼中的表达图式，进一步阐明SoxB1基因家族在脊椎动物神经系统发生过程中的作用。此外，我们还克隆了非洲爪蟾MGC85160基因并利用RT-PCR和胚胎整体原位杂交技术探测它在不同胚胎阶段的时空表达图式。结果表明母源性表达的MGC85160基因早期主要在动物极表达；从神经板期开始在发育的中枢神经系统和眼中表达，石蜡切片显示它主要在视网膜和晶状体中表达，说明该基因在爪蟾早期外胚层的模式化以及中枢神经系统的发育过程中可能起到重要作用。 此外，我们还研究了鱇浪白鱼的早期发育分期和眼睛特异基因的表达图式。鱇浪白鱼（Anabarilius grahami ）是云南抚仙湖的特有鱼种。我们首次完成了鱇浪白鱼早期发育的完整分期，主要包括合子期，卵裂期，囊胚期，原肠期，体节期和孵化期六个主要的时期。为了理解鱇浪白鱼眼睛的发育，我们克隆并检测了在眼发育早期起关键作用的基因Sox2, Pax6a, Six3a 和 Rx2的表达图式。结果表明这四个基因全部在尾芽期的前端神经板中表达，随后在视网膜原基细胞中表达明显。在晚期阶段，除Rx2外其它三个基因也在晶状体中表达。其表达模式与斑马鱼中同源基因的表达很相似，说明涉及眼发育的分子网络在鱇浪白鱼中也是高度保守的。

非洲爪蟾Brunol和MCPH基因在早期胚胎发育过程中表达的研究及随机定向酵母双杂交cDNA文库的构建

BRUNOL/CELF家族RNA结合蛋白在转录后调控（post-transcriptional regulation）中起着至关重要的作用，参与多种组织的发育过程。本研究中，我们描述了非洲爪蟾5个Brunol 基因的克隆与表达。其中只有Brunol2是母源性以及合子表达的，其它的4个Brunol基因都是合子表达的，但起始表达的发育时期有所不同。爪蟾发育过程中，Brunol1、4和5基因特异性地在神经系统中表达，包括脑、脊髓、眼泡和耳泡。Brunol2和3基因在体节中胚层与神经系统表达。Brunol2也在晶状体中有非常高的表达。在转染的Hela细胞中，BRUNOL1、2和3蛋白定位于细胞质和细胞核中，BRUNOL4和5只是定位于细胞质中，显示它们具有不同的功能。 人的microcephalin1（MCPH1）基因的遗传突变产生原发性小头症，而在人类的进化过程中，这个基因的变化可能对人脑体积的增加和认知能力的增强也起到重要的作用。但是对于MCPH基因在其它物种中功能的研究才刚刚开始。我们克隆了非洲爪蟾MCPH基因，发现非洲爪蟾具有A，B两个同源基因，其功能域的保守性较高，暗示非洲爪蟾MCPH基因仍然执行一些保守的功能，但MCPHB由于突变只编码一种截短的蛋白，目前尚不清楚它是否是有功能的。胚胎原位杂交的结果显示MCPHA，B基因在胚胎发育中的表达图式相似，但MCPHB的表达水平较低。在神经胚期，二者均表达于头部基板区，在尾芽期主要表达于咽鳃区，而在脑区的表达并不显著，与小鼠中的表达模式不同，提示在爪蟾中MCPH基因可能主要参与咽鳃区而不是脑的发育。 为了进一步筛选这些蛋白可能的结合因子，我们构建了非洲爪蟾双杂交cDNA文库。其中利用修饰的随机引物和特别设计的连接头在合成双链cDNA时，在下游合成一个SalI限制性酶切位点，可以将cDNA定向插入载体。通过对于空克隆率和插入片段长度等一系列参数的分析，表明这个定向cDNA文库的构建是成功的。

人类基因组中保守二级结构的纯净化选择及其在转录调控网络中的作用

保守序列是一种跨物种保守的基因组序列，而且绝大多数为非蛋白编码序 列。保守序列在人类遗传疾病中发挥着重要作用。其中，一部分保守序列能够 折叠形成二级结构。已鉴定的一些保守二级结构编码一些RNA 分子，如 microRNA、RNA 编辑序列和组蛋白mRNA 3’端非翻译区茎环结构等。但是，对 于绝大部分的保守二级结构，它们的生物学功能以及作用于它们上面的进化作 用力依然是未知的。 群体的SNP 数据在分析序列上的进化作用力时非常有效。SNP 在群体中的 频率会因为受到不同的进化作用力而表现出差异，而与其是否位于基因组中的 突变热点无关。对于受纯净化选择作用的SNP，它们的频率一般会比中性SNP 具有低的新生型等位基因频率（DAF）。我们运用生物信息学的方法，在人类基 因组保守二级结构中找到746 个SNP。这746 个SNP 与基因组其它区段的SNP 在突变模式上并不存在显著差异，在保守二级结构内同样存在突变热点。通过 与侧翼序列SNP 的分布比较发现，保守二级结构上SNP 密度约为其侧翼序列的 2/3。相比于侧翼序列SNP，有更高比例的保守二级结构SNP 具有低的DAF 值。 这些结果提示，有很多保守二级结构上的SNP 因为受到纯净化选择作用而在现 代人群中被剔除了。保守二级结构与侧翼序列在SNP 密度和DAF 上的差异要高 于保守序列与非保守序列之间的差异，提示保守二级结构是受到纯净化选择作 用最为严格的一类保守序列。我们发现，在保守二级结构内部，纯净化选择作 用的强度也有差异。茎区比环区具有更低的SNP 密度，而且有更高比例的茎区 SNP 具有低的DAF 值。这个结果提示，保守二级结构上的纯净化选择力主要作 用于茎区上的位点。我们推测，这可能是茎区上的突变往往比环区的突变对二级结构的造成更大的影响导致的。 我们通过寻找保守二级结构与转录因子SOX2、OCT4、NANOG、SUZ12 和C-MYC 结合位点之间的重叠，还分析了保守二级结构在转录调控网络中的作用。结果 显示，很多保守二级结构是作为转录因子的结合位点调控了许多与发育相关的 转录因子编码基因的表达。转录因子与保守二级结构之间的结合模式非常复杂， 可以有多个转录因子结合到同一个保守二级结构上，也可以是一个转录因子结 合到自身编码基因相关的保守二级结构上。不同的转录因子和保守二级结构结 合可以主导靶基因的特异模式，当绝大多数相关的保守二级结构与SUZ12 结合 时，基因表达受到抑制，而当绝大多数相关的保守二级结构不与SUZ12 结合时， 基因表达受到激活。在转录调控网络中，约有30%的保守二级结构是作为启动 子来调控基因的表达。因为转录因子SOX2、OCT4、NANOG、SUZ12 和C-MYC 仅仅 只结合到很小一部分保守二级结构上，提示可能还有更多的转录因子会结合到 保守二级结构上。因此，保守二级结构介导的转录调控网络要比目前已知的复 杂得多。

Molecular cloning, characterization and evolutionary analysis of phytoene desaturase (PDS) gene from Haematococcus pluvialis

Phytoene desaturase is one of the most important enzymes necessary for the biosynthesis of carotenoids in some cyanobacteria, green algae and plants. In this study, genomic DNA and cDNA of pds were cloned from unicellular green alga Haematococcus pluvialis strain323 using PCR and RT-PCR methods. The cDNA was cloned into plasmid pET-28a and efficiently expressed in Escherichia coli BL21. The complete genomic PDS gene of H. pluvialis, 3.3 kb in size, included eight exons and seven introns. To locate transcriptional regulation elements, an approximate 1 kb of 5'-flanking region was isolated by genome-walking method. Results of bioinformatic analysis showed several putative cis-elements e.g. the ABRE motif (abscisic acid responsive element), the C-repeat/DRE (dehydration responsive element) motif and the GCN4 motif were located in 5'-flanking region of pds. Results of phylogenetic analyses reveal that different sources of PDS genes form a separate clade, respectively, with 100% bootstrap support. Moreover, a maximum likelihood approach was employed to detect evidence of positive selection in the evolution of PDS genes. Results of branch-site model analysis suggest that 7.9% of sites along the green algal branch are under positive selection, and the PDS gene in green algae exhibits a different evolutionary pattern from its counterparts in cyanobacteria and plants.

Genes involved in complex adaptive processes tend to have highly conserved upstream regions in mammalian genomes

BACKGROUND:Recent advances in genome sequencing suggest a remarkable conservation in gene content of mammalian organisms. The similarity in gene repertoire present in different organisms has increased interest in studying regulatory mechanisms of gene expression aimed at elucidating the differences in phenotypes. In particular, a proximal promoter region contains a large number of regulatory elements that control the expression of its downstream gene. Although many studies have focused on identification of these elements, a broader picture on the complexity of transcriptional regulation of different biological processes has not been addressed in mammals. The regulatory complexity may strongly correlate with gene function, as different evolutionary forces must act on the regulatory systems under different biological conditions. We investigate this hypothesis by comparing the conservation of promoters upstream of genes classified in different functional categories.RESULTS:By conducting a rank correlation analysis between functional annotation and upstream sequence alignment scores obtained by human-mouse and human-dog comparison, we found a significantly greater conservation of the upstream sequence of genes involved in development, cell communication, neural functions and signaling processes than those involved in more basic processes shared with unicellular organisms such as metabolism and ribosomal function. This observation persists after controlling for G+C content. Considering conservation as a functional signature, we hypothesize a higher density of cis-regulatory elements upstream of genes participating in complex and adaptive processes.CONCLUSION:We identified a class of functions that are associated with either high or low promoter conservation in mammals. We detected a significant tendency that points to complex and adaptive processes were associated with higher promoter conservation, despite the fact that they have emerged relatively recently during evolution. We described and contrasted several hypotheses that provide a deeper insight into how transcriptional complexity might have been emerged during evolution.

Investigations into selective metabolic aspects of bifidobacteria: carbohydrate metabolism, fatty acid biosynthesis and plasmid biology

The gastrointestinal tract (GIT) is a diverse ecosystem, and is colonised by a diverse array of bacteria, of which bifidobacteria are a significant component. Bifidobacteria are Gram-positive, saccharolytic, non-motile, non-sporulating, anaerobic, Y-shaped bacteria, which possess a high GC genome content. Certain bifidobacteria possess the ability to produce conjugated linoleic acid (CLA) from linoleic acid (LA) by a biochemical pathway that is hypothesised to be achieved via a linoleic isomerase. In Chapter two of this thesis it was found that the MCRA-specifying gene is not involved in CLA production in B. breve NCFB 2258, and that this gene specifies an oleate hydratase involved in the conversion of oleic acid into 10-hydroxystearic acid. Prebiotics are defined as non-digestible food ingredients that beneficially affect the host by selectively stimulating growth and/or activity of one or a limited number of bacteria in the colon. Key to the development of such novel prebiotics is to understand which carbohydrates support growth of bifidobacteria and how such carbohydrates are metabolised. In Chapter 3 of this thesis we describe the identification and characterisation of two neighbouring gene clusters involved in the metabolism of raffinose-containing carbohydrates (plus related carbohydrate melibiose) and melezitose by Bifidobacterium breve UCC2003. The fourth chapter of this thesis describes the analysis of transcriptional regulation of the raf and mel clusters. In the final experimental chapter two putative rep genes, designated repA7017 and repB7017, are identified on the megaplasmid pBb7017 of B. breve JCM 7017, the first bifidobacterial megaplasmid to be reported. One of these, repA7017, was subjected to an in-depth characterisation. The work described in this thesis has resulted in an improved understanding of bifidobacterial fatty acid and carbohydrate metabolism, Furthermore, attempts were made to develop novel genetic tools.