|Home | About | Journals | Submit | Contact Us | Français|
The small Nhp6 protein from budding yeast is an abundant protein that binds DNA non-specifically and bends DNA sharply. It contains only a single HMGB domain that binds DNA in the minor groove and a basic N-terminal extension that wraps around DNA to contact the major groove. This review describes the genetic and biochemical experiments that indicate Nhp6 functions in promoting RNA pol III transcription, in formation of preinitiation complexes at promoters transcribed by RNA pol II, and in facilitating the activity of chromatin modifying complexes. The FACT complex may provide a paradigm for how Nhp6 functions with chromatin factors, as Nhp6 allows Spt16-Pob3 to bind to and reorganize nucleosomes in vitro.
Saccharomyces cerevisiae has seven genes expressing HMGB proteins : HMO1, NHP10, ABF2, ROX1, IXR1, NHP6A, and NHP6B. NHP6A and NHP6B encode highly homologous proteins of 93 and 100 amino acids in length. The Nhp6A and Nhp6B proteins differ significantly at their N-termini, but over the 90 amino acid core region they are 89% identical and 96% similar .
Nhp6A is an abundant protein, present at 50,000 to 70,000 molecules per haploid cell . This corresponds to one Nhp6A molecule for every 1–2 nucleosomes, which is similar to the value of 1 HMG1/2 per 3 nucleosomes reported for mammalian cells . ChIP-chip experiments suggest Nhp6 localization in the vicinity of transcription start sites parallels that of nucleosomes [K. Yen and B.F. Pugh, personal communication; 5]. The Nhp6A protein is present at a concentration between three times  to ten times  higher than Nhp6B, consistent with higher Nhp6A transcript levels . Expression of NHP6 genes is regulated by the concentration of Nhp6 protein or RNA, since overexpression of Nhp6B protein results in a dramatic decrease in NHP6A expression . Additionally, overproduction of Nhp6 is toxic to cells .
Like most HMGB proteins, Nhp6A binds DNA in a sequence-nonspecific fashion [8, 9]. Gel shift experiments show multiple Nhp6A complexes on a 98 bp DNA fragment . Nhp6A binds to DNA as a monomer in a stepwise manner with an affinity of 1–10 nM for the initial complex [10–12]. Nhp6A bends DNA sharply, as shown by the ability to supercoil circular DNA , and by ligase-mediated circularization assays where Nhp6A promotes circularization of DNA molecules as short as 66 bp . Additionally, Nhp6A binds more tightly and more stably to curved DNA than to linear DNA .
The structure of Nhp6A has been solved by NMR [11, 14]. Nhp6 adopts the typical L-shaped HMGB fold, both free in solution and when complexed with DNA. The inside of the L binds to DNA in the minor groove resulting in a 70° bend in DNA, similar to the sharp DNA bend seen in other HMGB-DNA structures. Specific Nhp6A side chains intercalate into the DNA minor groove, and studies of Nhp6A bound to cisplatin-modified DNA suggest these intercalating residues are important in inducing the bend in DNA . The N-terminal region (amino acids 8–16) of the protein has numerous basic residues and this region wraps around the DNA to make contacts within the narrowed major groove. Mutations within this basic N-terminal region affect Nhp6 function in vivo . Studies with HMGB chimeras containing the basic N-terminal region of Nhp6A show this region enhances DNA flexibility .
Yeast cells with a gene disruption of a single NHP6 gene, either NHP6A or NHP6B, are viable and have no visible growth defects. In contrast the nhp6a nhp6b double mutant grows slowly at 30°C and does not grow at all at 38°C . Thus Nhp6 is non-essential, but is important for normal growth. While the Nhp6A and Nhp6B proteins are similar and the nhp6a and nhp6b single mutants grow normally, growth competition experiments [17, 18] show that nhp6a mutants have a subtle but reproducible growth defect compared to wild type strains, while nhp6b mutants compete well with wild type (T. Engar and D.J.S., unpublished results). Cells shifted to the non-permissive temperature display morphological defects including elongated buds and enlarged necks . The temperature sensitive growth defect of nhp6a nhp6b double mutants is suppressed by the presence of 1M sorbitol in the medium; this osmoremediation did not suppress the morphological defects, however.
Some of the residues most highly conserved among HMGB proteins were mutated in Nhp6A and tested for function both in vivo and in vitro . Surprisingly, most of the changes allowed Nhp6A to function normally in supporting growth at 38°C in vivo, and were normal in binding and bending DNA. Interestingly, the F31S and P44L mutations were defective for in vivo functions, but in vitro studies showed these mutant proteins could bind and bend DNA as well as wild type. Based on these results the authors suggest that the biological function of Nhp6 requires interaction with other proteins, a topic that is explored in subsequent sections. Nuclear entry by Nhp6 is via a pathway independent of the Ran import pathway, but requires calmodulin [20, 21].
Systematic analysis of the yeast knockout collections  has provided a substantial insight as to phenotypes, as have the use of synthetic genetic analysis  to identify genetic interactions. This approach does not work well for nhp6 mutations, as nhp6a and nhp6b single mutants do not display significant phenotypes, and the nhp6a nhp6b double mutant (here referred to as nhp6ab) usually must be used in genetic analyses. Studies have shown that nhp6ab mutants are sensitive to nitrogen starvation , are slightly more resistant to hydrogen peroxide and ultraviolet irradiation , and are defective for galactose growth in the S288c strain background . nhp6ab mutants have a weak Spt-phenotype and are sensitive to the transcription elongation inhibitor 6-AU [25, 26].
A role for Nhp6 in chromatin is seen in genetic studies (Table 1), as a number of chromatin regulators show synthetic lethality or synthetic sickness when combined with nhp6ab, including the GCN5 histone acetyl transferase , the Swi/Snf and RSC ATP-dependent chromatin remodelers [26–29], the SET2 histone methyltransferase , the SPT4 and SPT5 elongation factors [25, 26], the REG1 repressor , the NHP10 HMGB factor , subunits of the Ccr/Not , Ino80 , Paf1 , NuA4 , SAGA , and FACT complexes [25, 26], as well as mutations affecting histone tails . nhp6ab mutants also show synthetic defects when combined with mutations in basal transcription factors such as TFIIA . TBP [34, 35], and MOT1, a TBP interacting factor . Additive defects are also seen when nhp6 mutations are combined with promoter mutations in the pol III-transcribed SNR6 gene  and with RPC40, a subunit of pol I and pol III . Genetic links are also provided by the genes whose defects can be suppressed by NHP6 overexpression, including mutations in the RSC and FACT chromatin factors [25, 28], and the SWI6 subunit of the SBF DNA-binding factor . NHP6 overexpression suppresses a TBP SWI2 double mutant suggesting a role in preinitiation complex formation , and also suppresses a slk1-1 spa2Δ double mutant suggesting a role in MAP kinase cascades . Nhp6 therefore has global effects in a range of processes linked to appropriate regulation of transcription through maintenance of normal chromatin structure.
Studies have shown a critical role for Nhp6 in transcription by RNA polymerase III, which transcribes tRNAs, 5S RNA, and other small RNA molecules. A screen was conducted to identify multicopy plasmids that can suppress the nhp6ab growth defect at 37°C, and this screen identified BRF1 and SNR6 . BRF1 encodes the limiting component of TFIIIB, one of the basal factors required for pol III transcription, and SNR6 is a pol III-transcribed gene that encodes the U6 spliceosomal RNA. Another genetic screen used a SNR6 gene with a promoter mutation causing inefficient expression and growth defects . These authors screened for multicopy suppressors and identified plasmids with NHP6A, NHP6B, and the limiting BRF1 factor . Expression of U6 snRNA in vivo is reduced in nhp6ab mutants, and Nhp6 can stimulate pol III transcription in vitro in studies with purified systems [38, 39]. Thus Nhp6 plays an important role in SNR6 expression, and the nhp6ab temperature sensitive phenotype is due, at least in part, to decreased SNR6 expression. The nhp6ab defect can be suppressed by increasing the number SNR6 templates, and both the nhp6ab defect and the SNR6 promoter mutations can be suppressed by overexpressing BRF1, the limiting component. Combining nhp6ab mutations with mutations in the SNR6 promoter, either at the 5' TATA element or the 3' flanking A and B block region, results in lethality . MNase digestion experiments show that there is a specific chromatin structure at the SNR6 TATA region, and this is altered in an nhp6ab mutant . TBP is a subunit of TFIIIB that binds at the 5' flanking TATA element of pol III-transcribed genes, and chromatin immunoprecipitation (ChIP) experiments show reduced TBP binding to the SNR6 promoter in nhp6ab mutant . Nhp6 provides transcriptional initiation fidelity to tRNA genes in a highly purified in vitro system, and in vivo studies show changes in transcriptional start sites at tRNA genes in an nhp6ab mutant . The RSC chromatin remodeler interacts with Nhp6 , and global ChIP-chip studies show high RSC occupancy at pol III-transcribed genes . Finally, tRNA genes can provide a barrier to the spread of heterochromatin, and this barrier function is compromised in an nhp6ab mutant .
Strains with mutations in both NHP6A and NHP6B display significant defects in transcriptional activation by RNA polymerase II. Paull et al.  demonstrated that activation of a number of inducible promoters is reduced in nhp6ab mutants, including copper induction of CUP1, low glucose induction of CYC1, DNA damage induction of DDR2, and galactose induction of GAL1 . The GAL1 defect is consistent with the nhp6ab defect in growth on galactose media in some strain backgrounds . The nhp6ab mutations did not affect PHO5, and thus not all inducible genes are affected . It was subsequently shown that nhp6ab mutants are also defective for induction of the CHA1 gene by serine , the SUC2 gene by sucrose , and the FRE2 gene by low iron . Chromatin structure at the CHA1 locus was altered in the nhp6ab strain . My laboratory showed that expression of the HO gene is reduced in an nhp6ab mutant , consistent with identification of NHP6 as a multicopy suppressor that allows HO expression despite a mutant Swi6 activator . While HO is not induced by a metabolite, its expression is tightly cell cycle regulated. Genetic analysis showed that the nhp6ab defect in HO expression could be suppressed by a number of mutations, including sin3 (which affects the Rpd3(L) histone deacetylase), spt3 (a component of the SAGA histone or protein acetyltransferase complex), and sin4 (a component of the Pol II Mediator complex) . These latter results suggest Nhp6 functions at the chromatin level in regulating gene expression, an idea supported by the observation that a multicopy plasmid with NHP6 allows HO expression in the absence of the normally required GCN5 histone acetyltransferase . Microarray experiments conducted to examine gene expression globally have not identified major defects in gene expression in nhp6ab mutants [P. Eriksson and D.J.S, unpublished; 43], although modest effects of the nhp6ab mutations were observed on expression of methionine biosynthetic genes, pheromone-regulated genes, and genes involved in the mating response were seen . These results are consistent with the interpretation that Nhp6 plays an important supportive role in global regulation of transcription rather than an essential role at a subset of genes.
Genetic and biochemical experiments both suggest that one function of Nhp6 is to facilitate preinitiation complex formation at promoters. In vivo studies with chimeric promoters suggested that Nhp6 has effects at both the UAS and core regions of promoters . In vitro binding experiments showed that Nhp6A promotes TFIIB to a DNA-TBP-TFIIA complex in vitro , and that Nhp6A can facilitate the in vitro interaction of TBP with DNA, especially in the presence of TFIIA .
A series of genetic studies provide strong support for the idea of interactions between Nhp6 and basal pol II factors. Overexpression of TBP suppresses the nhp6ab defect in HO expression as well as the nhp6ab temperature sensitive growth defect . Strains with an epitope tag on TBP at either the N- or C-terminus are viable and healthy, but these TBP-tags are lethal in an nhp6ab mutant . This result led to a screen to identify point mutations in TBP that are viable on their own, but lethal in combination with nhp6ab . Interestingly, the TBP mutations identified this way cluster on the protein surface and particularly affect residues known to interact with the TFIIB and TFIIA basal factors. Many of these TBP alleles are also lethal when combined with a mutation in either the SWI2 subunit of Swi/Snf or the GCN5 histone acetyltransferase subunit of SAGA . Some of these synthetic lethal interactions can be suppressed by multicopy plasmids with either TFIIA or NHP6A . nhp6ab is also lethal when combined with TFIIA mutations that affect its interaction with TBP that are otherwise tolerated .
There are also genetic interactions between NHP6 and factors known to interact with TBP. The Spt3 subunit of the SAGA complex shows allele-specific interactions with TBP , and an spt3 mutation also suppresses both the nhp6ab defect in HO expression and the temperature sensitive growth defect . An spt3 mutation suppresses some of the nhp6ab TBP synthetic lethal interactions , as well as the synthetic lethality between nhp6ab and the SPT16 subunit of FACT . Mot1 also interacts with TBP [47, 48], and mot1 mutations are synthetically defective with nhp6ab . This mot1 nhp6ab synthetic lethality can be suppressed by multicopy plasmids containing the genes for TBP or TFIIB . Similarly, nhp6ab is synthetically defective with a ccr4 mutation affecting the Ccr4/Not complex, and this can be suppressed by a multicopy plasmid with TBP .
One study detected an interaction between Nhp6 and Mediator using the split-ubiquitin two hybrid system, and showed that overexpression of the Rpb4 subunit of RNA pol II suppressed the 6-AU sensitivity of nhp6ab mutants . The 6-AU sensitivity of a nhp6ab mutant can also be suppressed by an spt3 mutation or by multicopy SNR6 .
These experiments all support the idea that Nhp6 works in the same pathway as the Swi/Snf chromatin remodeling complex and the Gcn5 histone acetyltransferase in promoting transcription by facilitating the interaction of TBP and TFIIA on promoter DNA (Fig 1). The synthetic lethal interactions are consistent with this idea: While gene deletions eliminating SWI2, GCN5 or NHP6AB are tolerated, inactivating any two of these pathways, in the swi2 gcn5, the swi2 nhp6ab, or the gcn5 nhp6ab mutant, results in synthetic lethality. Mutations that compromise TBP or TFIIA interaction can be tolerated, but are lethal when NHP6, SWI2 or GCN5 are mutated.
Nhp6 has been shown to interact with a variety of chromatin modifiers, including FACT, Swi/Snf, RSC, Ssn6, and Spt6 (Table 1). By far the most work has been done on how Nhp6 stimulates FACT activity, and I will discuss FACT first. The FACT chromatin reorganizer plays important roles in DNA replication, transcription elongation, and transcription initiation [for review see 50]. Recent work has shown that FACT is required for redeposition of histones evicted by elongating RNA polymerase  and for nucleosome eviction at certain promoters [52–55]. In vitro studies show FACT changes the accessibility of DNA within nucleosomes without repositioning the histone octamer core relative to the DNA [12, 25, 52]. FACT affects chromatin structure without hydrolyzing ATP, and thus FACT functions quite differently from the Swi/Snf family of ATP-dependent chromatin remodeling factors.
FACT is a heterodimer composed of two subunits, Spt16-Pob3 in yeast, and Spt16-SSRP1 in metazoans. The Spt16 protein is conserved along its entire length from yeast to humans. In contrast, while Pob3 is conserved over most of its length, the fungal forms lack the HMGB motif present at the C-terminus of the metazoan homologs. The evidence suggests that the Nhp6 HMGB protein provides a DNA-binding (or DNA-bending) activity for the yeast Spt16/Pob3 complex. The Spt16-Pob3 heterodimer does not bind to DNA or nucleosomes in vitro; addition of Nhp6, however, permits the formation of a FACT – nucleosome complex . The structure of nucleosomes, as assayed by sensitivity to DNaseI and restriction enzymes, is altered by FACT (Spt16-Pob3-Nhp6) [25, 52]. Nhp6 does not stably associate with Spt16-Pob3 , although an association has been described using an immunoprecipitation protocol conducted at 50 mM potassium acetate .
Careful studies examining the amount of Nhp6 required to affect nucleosome structure in vitro have been informative . Adding increasing concentrations of Nhp6 to in vitro binding reactions results in discrete complexes that differ by one Nhp6 monomer, whether the DNA is free in solution  or present in a nucleosome . When sufficient Nhp6 is present the nucleosomes become sensitive to DNaseI cleavage or hydroxyl radical damage and also can bind Spt16-Pob3 [12, 52]. Importantly, these changes all occur at similar Nhp6 concentrations, suggesting the effects are related. Additionally, the concentration of Nhp6 needed for these effects is 10-fold higher than that needed for simple Nhp6 binding. It has been suggested that HMGB proteins function as "shape chaperones," repeatedly binding to and releasing from DNA and thus allowing the DNA to form other shapes [56, 57]. Thus Nhp6 may repeatedly bind to nucleosomal DNA and thus promote formation of an altered nucleosome as assayed by nuclease sensitivity, and this altered nucleosome is then receptive to Spt16-Pob3 binding. The stoichiometry of Nhp6, compared to Spt16-Pob3, shows that multiple Nhp6 molecules are needed for a FACT-nucleosome complex to form. The authors propose that nucleosome reorganization by FACT is a two step process (Fig 2), with the concerted action of multiple Nhp6 molecules converting nucleosomes as the first step, and the second step requiring Spt16-Pob3 .
The biochemical analysis suggests Spt16-Pob3 require Nhp6 for activity, and there is also strong genetic support for this idea. SPT16 and POB3 mutations are synthetically lethal with nhp6ab, and Nhp6 overexpression suppresses mutations in these FACT subunits [25, 26]. SPT16 and POB3 are both essential genes, raising the question of how an nhp6a nhp6b double mutant is viable. It is possible that other proteins besides Nhp6 can provide the necessary function for Spt16-Pob3 in vivo. Candidates for this activity would be other HMGB proteins such as Hmo1 and Nhp10; however, the nhp6a nhp6b hmo1 nhp10 quadruple deletion strain is viable . Thus Spt16-Pob3 might be able to reorganize nucleosomes in vivo without Nhp6, albeit less efficiently, or possibly using other types of DNA-binding proteins.
It seems likely that Nhp6 works with other chromatin modifiers besides FACT. Nhp6 overexpression suppresses mutations in subunits of the Swi/Snf and Rsc ATP-dependent chromatin organizers [27, 28], and there are synthetic interactions between nhp6ab and either Swi/Snf or Rsc [26–29]. It is intriguing that mammalian Swi/Snf contains an HMG domain that is absent in yeast Swi/Snf , a situation analogous to mammalian and yeast FACT. In vitro experiments show RSC stimulates Nhp6 binding to nucleosomes , and Nhp6 stimulates nucleosome sliding by RSC and Swi/Snf (M. Hepp, X. Steinberg, and J.L. Gutierrez, personal communication). Nhp6 also promotes nucleosome binding in vitro by the Spt6 histone chaperone (H. Xin and T. Formosa personal communication), analogous to the loading of Spt16-Pob3 by Nhp6. Finally, genetic experiments suggest Nhp6 facilitates recruitment of the Ssn6 coactivators to the FRE2 genes . This suggests that Nhp6 alters nucleosome structure globally, and this affects the activity of many factors that act upon nucleosomes. The complex genetic interactions observed could be explained solely by an indirect effect in which failure of Nhp6 to maintain a permissive nucleosome structure forces a broad range of factors that must overcome nucleosomal barriers to act with full efficiency. Alternatively, The local concentration of Nhp6 could be increased by specific protein-protein interactions like the weak binding detected with Spt16-Pob3 . Finally, it is intriguing that mammalian HMGB facilitates nucleosome sliding by the ACF/CHRAC remodeling factor , analogous to the activity of Nhp6.
A recent paper shows that the Msh2-Msh6 DNA mismatch repair factor stimulates binding of Nhp6 to DNA in vitro . Additionally, Nhp6 reduces the nonspecific binding of Msh2-Msh6 to homoduplex DNA, but Nhp6 does not affect Msh2-Msh6 binding to mismatched DNA. Thus Nhp6 modulates the binding activity of the Msh2-Msh6 DNA mismatch repair complex towards targets with heteroduplex DNA.
The 11 kD Nhp6 protein is small in size, but is required for normal growth of budding yeast. It is an abundant protein that binds DNA non-specifically and bends DNA sharply. Genetic experiments suggest Nhp6 interacts with variety of chromatin modifiers, including FACT, Swi/Snf, RSC, and Ssn6. Biochemical experiments with FACT suggest that Nhp6 bends the DNA within a nucleosome, allowing recruitment of Spt16-Pob3 and then reorganization of a nucleosome as assayed by nuclease sensitivity. Preliminary biochemical work indicates Nhp6 can affect the activity of the Swi/Snf and RSC remodelers and of the Spt6 histone chaperone in vitro, perhaps through a similar mechanism. Nhp6 is also required for efficient transcription by RNA polymerase III, particularly at the SNR6 gene, and also for formation of the preinitiation complex pol II basal factors. It is possible that these latter effects are indirect, due to the nhp6ab mutation affecting the structure of nucleosomes or the activity of chromatin complexes such as FACT.
I thank Tim Formosa, Jose Gutierrez, and Frank Pugh for granting permission to mention unpublished results. I especially thank Tim Formosa for many helpful discussions and for comments on the manuscript. This work was supported by grants R01-GM39067 and R01-GM64649 from the National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.