In this work we have shown that the Spt16-M domain is an important player in functional interactions between Spt16 and histone H3. Mutations within the Spt16-M domain suppress several mutant growth phenotypes conferred by the histone H3 mutant H3-L61W, including growth defects at low temperatures and in the presence of several drugs. At the molecular level, many Spt16-M domain mutations suppress the Spt16 3′-accumulation phenotype at the PMA1 gene and all Spt16-M domain mutations suppress an intragenic cryptic transcription initiation phenotype seen in H3-L61W cells. The extreme C-terminus of Spt16 is also implicated in functional interactions with histone H3 since two small C-terminal truncations of Spt16 were also found to suppress H3-L61W growth and intragenic cryptic transcription initiation phenotypes. Finally, in experiments carried out in the context of wild-type histone H3, we have provided evidence in support for roles for the Spt16-M domain in ensuring proper distribution of Spt16 across transcribed genes and in preventing aberrant cryptic transcription initiation.
The observation that the Spt16-M domain mutants suppress the Spt16 3′-accumulation phenotype at PMA1
seen in H3-L61W cells points to a role for this domain in regulating the departure of Spt16 following transcription. However, in the context of H3-WT, the levels of Spt16 detected at the 3′ region of PMA1
were comparable between wild-type Spt16 cells and cells expressing any of the Spt16 mutants (see green bar-graphs in ). Therefore, the Spt16 mutants do not suppress the Spt16 3′-accumulation phenotype seen in H3-L61W cells simply by an inherent and general ability of these mutants to more readily disengage from chromatin following transcription. Instead, we envision that the Spt16 suppressors may render Spt16 more sensitive to a putative Spt16-departure signal. We have previously proposed a model in which Spt16 normally requires a signal – perhaps a post-translational modification of one or more core histone residues – to be able to depart chromatin following transcription 
. In the context of the H3-L61W mutation, this signal is either not properly initiated and/or propagated throughout the nucleosome, thereby resulting in a crippled signal that, in turn, leads to accumulation of Spt16 at 3′ ends of genes. If the Spt16-M domain mutants we have isolated are indeed more sensitive to the Spt16-departure signal, they would be expected to behave as shown in our studies: they would show more efficient departure from a gene's 3′ end in situations where the departure signal is impaired (i.e.
in the context of H3-L61W nucleosomes), but they would have no effect in situations where the signal is fully operational (i.e.
in the context of H3-WT nucleosomes). Thus, our results point to a role for the Spt16-M domain in regulating Spt16 dissociation from chromatin following the transcription process and are consistent with a model in which the Spt16-M domain serves as a sensor for a putative Spt16-departure signal.
Our studies show that the Spt16-M domain mutants can suppress the H3-L61W cryptic transcription initiation phenotype at the FLO8:HIS3
reporter gene when cells are grown in glucose-containing medium while causing a cryptic transcription initiation phenotype at the same gene in the context of H3-WT when cells are grown in galactose-containing medium. These results can be explained by a model of allele-specific Spt16-histone H3 interactions. In this model, the Spt16-M domain and histone H3 interact directly to promote nucleosome reassembly during transcription elongation. The H3-L61W mutation disrupts this interaction, leading to impaired nucleosome reassembly and cryptic intragenic transcription. The Spt16-M domain mutants restore the interaction with H3-L61W through compensatory structural changes leading to more efficient nucleosome reassembly and, as a result, suppression of cryptic transcription. These structural changes, however, decrease the affinity between the Spt16-M domain and H3-WT, resulting in defective nucleosome reassembly and cryptic transcription in H3-WT cells. The Spt16-M domain may interact with histone H3 through the histone H3 N-terminal tail since this tail has been shown to interact with S. cerevisiae
yFACT in vitro
through yFACT regions other than the Spt16-NTD domain 
. The observation that the Spt16 mutants show suppression of H3-L61W-dependent cryptic transcription initiation at FLO8-HIS3
only in conditions in which the reporter gene is expected to be largely inactive seems puzzling at first given the fact that cryptic intragenic transcription is thought to be dependent on active transcription originating from a bona fide
upstream promoter. However, recent studies have suggested that even at very lowly transcribed genes and at genes expected to be inactive, transcription-dependent nucleosome disassembly and Spt16-mediated reassembly can still occur 
. Therefore, suppression of the H3-L61W-dependent FLO8-HIS3
cryptic transcription phenotype by the Spt16 mutants is likely to be due to an increased ability of the Spt16 mutants to reassemble H3-L61W nucleosomes during the transcription elongation process. We do not rule out, however, alternative models that could account for cryptic transcription at FLO8-HIS3
in H3-L61W cells grown in glucose-containing medium and for the mechanism of suppression by the Spt16 mutants. For example, it is possible that an abnormally open chromatin structure permissive to intragenic cryptic transcription is generated during DNA replication in the context of H3-L61W and that the Spt16 mutants suppress cryptic transcription by promoting a more compact chromatin structure on the newly replicated DNA molecules. Alternatively, expression of HIS3
in the context of the FLO8:HIS3
reporter in H3-L61W cells may be a result of changes in chromatin structure over the cryptic promoter element of FLO8
that allows for transcription initiation to occur in the absence of any transcription originating from the GAL1
promoter. In this scenario, the Spt16 mutants would suppress cryptic transcription by reestablishing a closed chromatin structure over the cryptic promoter element through activities related to its role in transcription initiation. Our results lay the foundations for future experiments, such as in vitro
biochemical assays, that will further explore the significance of the interactions between the Spt16-M domain and histone H3 in regulating chromatin dynamics.
We have entertained the possibility that the abilities of Spt16 to reassemble nucleosomes during transcription elongation and to depart 3′ ends of genes following transcription are functionally related to each other. For example, it could be envisioned that in order to properly depart chromatin following transcription, Spt16 needs to efficiently reassemble the last nucleosome at the 3′ end of a transcribed gene. However, this hypothesis is not supported by our data. For example, in the context of H3-L61W nucleosomes, Spt16-E735G is among the strongest suppressors of the Spt16 3′-accumulation phenotype () but is one of the weakest suppressors of the cryptic transcription initiation phenotype (). Thus, at least in this case, more efficient departure from a gene's 3′ end does not appear to correlate with a significant increase in nucleosome reassembly activity.
Our results also point to a role for the extreme C-terminus of Spt16 in directing functional interactions with histone H3. The two C-terminal truncation mutants isolated in our screens, Spt16-E989stop and Spt16-E1004stop, are predicted to lack the last 47 and 32 residues of Spt16, respectively, and western blot experiments confirm that these proteins are smaller in size compared to wild-type Spt16 (see ). The fact that these mutants are able to support life indicates that Spt16 can withstand small C-terminal truncations and still provide essential functions. In an H3-L61W background, the truncation mutants partially suppress the cryptic transcription initiation phenotype at FLO8 (see ) but do not suppress the Spt16 3′-accumulation phenotype at PMA1 (see ). In H3-WT cells, the truncation mutants allow for low levels of cryptic transcription initiation in the (hht1-hhf1)Δ background (see ). These data suggest that the extreme C-terminal region of Spt16 is involved in the prevention of cryptic intragenic transcription and raise the possibility that the extreme C-terminus of Spt16 may influence histone deposition during transcription elongation. This region of Spt16 might exert its effects directly through interactions with histone H3 or indirectly by ensuring the structural and functional integrity of the Spt16-M domain. Alternatively, it is possible that the extreme C-terminal region of Spt16 is involved in recruitment of other factors involved in interactions with histone H3.
Our screens for Spt16 mutants able to suppress H3-L61W phenotypes were not saturated. Therefore, it is likely that mutations in residues of Spt16 other than the ones we have reported in this and our previous work can display genetic interactions with the H3-L61W mutation. An inspection of three additional Spt16 suppressors we isolated in our screens provides some relevant insights in this regard. These three mutants – Spt16-K456I, Spt16-S625L and Spt16-S841W – were found to suppress the Cs- phenotypes of H3-L61W cells but were not further analyzed because each also contained a second mutation within the region targeted for the PCR mutagenesis (G3376A, C-390A and T-182Δ, respectively – see to locate these positions) that could potentially have an effect on SPT16 expression levels and therefore could not be ruled out for being responsible, at least in part, for the suppression phenotypes (the only mutants with two nucleotide changes across this region that were included in the study were those in which one of the two mutations resulted in a silent mutation within the SPT16 coding sequence, see ). However, these three mutants do encode Spt16 proteins with single amino acid changes and it is therefore likely that these mutations are in fact the ones responsible for the suppression of the H3-L61W Cs− phenotype. Whereas Spt16-G841W harbors a mutation within the Spt16-M domain, thereby supporting the notion that the Spt16-M domain is functionally related to histone H3, the Spt16-K456I mutation is located in a region connecting the Spt16-NTD and Spt16-D domains and the Spt16-S625L mutation is located in the Spt16-D domain. Thus, these regions may also be involved in interactions with histone H3, although the ability of Spt16-K456I or Spt16-S625L to suppress the H3-L61W-dependent Spt16 3′-accumulation phenotype and cryptic transcription initiation defect has not been assessed. That regions other than the Spt16-M domain and the extreme C-terminus might be involved in functional interactions with H3-L61W is supported by the observation that the Spt16-P599Q mutation, which is located in the Spt16-D domain, can also suppress H3-L61W phenotypes, although it is possible that this mutation and the Spt16-S625L mutation described above, due to their proximity to the Spt16-M domain, exert their effects indirectly by compromising the integrity of the Spt16-M domain. In any case, the finding that many mutations in the Spt16-M domain and two C-terminal truncations show suppression of several H3-L61W phenotypes highlights the importance of these two regions in functional interactions with histone H3. The fact that, despite the unbiased nature of the screens we carried out, a large proportion of the suppressive mutations are clustered within the Spt16-M domain further underscores the relevance of this domain in histone H3-dependent functions.