A current view on long noncoding RNAs in yeast and filamentous fungi

Long noncoding RNAs (lncRNAs) are crucial players in epigenetic regulation. They were initially discovered in human, yet they emerged as common factors involved in a number of central cellular processes in several eukaryotes. For example, in the past decade, research on lncRNAs in yeast has steadily increased. Several examples of lncRNAs were described in Saccharomyces cerevisiae and Schizosaccharomyces pombe. Also, screenings for lncRNAs in ascomycetes were performed and, just recently, the first full characterization of a lncRNA was performed in the filamentous fungus Trichoderma reesei. In this review, we provide a broad overview about currently known fugal lncRNAs. We make an attempt to categorize them according to their functional context, regulatory strategies or special properties. Moreover, the potential of lncRNAs as a biotechnological tool is discussed.


Introduction
Pervasive transcription including intergenic and antisense regions evolved as a common feature in higher and lower eukaryotes (David et al. 2006;Dutrow et al. 2008;Nagalakshmi et al. 2008;Wilhelm et al. 2008). A very heterogeneous group of noncoding transcripts are the long noncoding RNAs (lncRNAs). They are distinguished from small RNA species upon their size of > 200 nt Kapranov et al. 2007), rather than upon any other property. The presence of a poly(A)-tail as well as a 5′-methylguanosine cap or posttranscription splicing events are optional features Kapranov et al. 2010;Yin et al. 2012). More common characteristics of lncRNAs are poor conservation and expression at low levels compared to protein encoding genes Djebali et al. 2012). Furthermore, lncRNAs often fold into complex, high ordered structures (Mercer and Mattick 2013). Transcripts targeted by degradation are also assigned to the lncRNAs (Schulz et al. 2013;van Dijk et al. 2011;Wyers et al. 2005;Yin et al. 2012). Anyhow, it should be considered that the discrimination between lncRNAs and coding transcripts is vague. Several not annotated transcripts, which were initially predicted to lack a protein encoding capacity, were found to be associated with the polyribosomes as they contain small open reading frames . The modes of action and also the processes, in which lncRNAs are involved in, are manifold. Their strategies range from physical interference with the transcription of adjacent or overlapping sense or antisense-oriented target genes in cis, over nucleosome repositioning, histone modifications or the recruitment of chromatin remodelling factors, to trans interactions with proteins, DNA or other RNAs, in some cases providing a scaffold for the attachment of multiple factors (reviewed in Ponting et al. 2009). Predominantly, lncRNAs act as repressors, yet also examples for positive regulation of the target genes have been reported (Krishnan and Mishra 2014).
Most lncRNAs were described in human and other mammals Fantom Consortium 2005;Hon et al. 2017;Iyer et al. 2015). In the last years, also the number of lncRNAs identified in the well-studied budding yeast Saccharomyces cerevisiae and model fission yeast Schizosaccharomyces pombe has rapidly increased (reviewed in (Niederer et al. 2017;Yamashita et al. 2016). In ascomycetes, such as Neurospora crassa and Magnaporthe oryzae, the physical presence of lncRNAs was reported (Cemel et al. 2017;Jain et al. 2017). And just recently, the first functional characterization of a lncRNA in the filamentous fungus Trichoderma reesei was achieved (Till et al. 2018). In this review, we give an overview about relevant research on fugal lncRNAs. Attempts for categorizing them are discussed, and outstanding examples are presented in more detail.

The challenge of classifying lncRNAs
In contrary to protein encoding genes, lncRNAs do not form large homologous families (Ponting et al. 2009). Thus, the attempt to classify of lncRNAs is a challenging task. Examples of thoroughly described lncRNAs identified in different fungi are listed in Table 1. The categorization was made on a cis or trans mode of action, the functional context, regulatory mechanism or special properties. Anyway, there are hardly any overlaps of the groups formed by the different strategies of classification. Different regulatory mechanisms are followed by lncRNAs involved in the same category of cellular processes, and also, special properties are not necessarily assigned to a functional context. Also, a cis or trans mode of action is not consistent with the function or mechanism. Conclusively, at least at the current state of knowledge, a generally valid categorization of fungal lncRNAs is rather pointless. Instead, we favour grouping by means of different criteria, consequently tolerating multiple assignments of some fungal lncRNAs.
Cis and trans acting lncRNAs Most lncRNAs described in yeast act in cis. This means that they regulate the expression of genes at proximal locations on the same chromosome, whereas trans-acting lncRNA affect either distal loci or the same locus but located on a homologue chromosome. Some cis-acting lncRNAs were demonstrated to act strictly in cis because trans-expression in diploids or ectopic expression in a lncRNA-deleted background results in a loss of function of the lncRNA. Among those are the CDC28 antisense lncRNA (Nadal-Ribelles et al. 2014), prt (Shah et al. 2014), pHO-lncRNA (Yu et al. 2016), ICR1 and PWR1 (Bumgarner et al. 2009) and RME2 and RME3 (Hongay et al. 2006). In contrast to this, for some lncRNAs initially described as cis-acting factors, the function was shown to be position independent. One example for this is the PHO84 antisense transcript, which belongs to the group of cryptic unstable transcripts (CUTs) and triggers silencing of its sense gene PHO84 in S. cerevisiae in aged cells (Camblong et al. 2009;Camblong et al. 2007).
Examples for per se trans-acting lncRNAs in fungi are rare. In the budding yeast S. cerevisiae, a trans-mechanism was stated for the stable unannotated transcript (SUT) SUT457, which physically interacts with 12 genes essential for telomere organization and homeostasis (Kyriakou et al. 2016). Another example is SPNCRNA.1164, a regulator of atf1 expression in response to oxidative stress in S. pombe (Leong et al. 2014). Furthermore, recently, a lncRNA termed HAX1 was identified as a trans-activator of cellulase expression in T. reesei (Till et al. 2018). Interestingly, all three of these lncRNAs (i.e. SUT457, SPNCRNA.1164 and HAX1) have a positive regulatory impact on their target genes. However, also yeast lncRNAs with a repressing function in trans are known. For example, members of the class of the Xrn1 unstable transcript (XUTs), such as TY1, frequently mediate silencing of translocated target genes (Berretta et al. 2008).

The role of lncRNAs in cellular processes
Some lncRNAs can be categorized according to their functional roles in the yeast cells. IRT1, RME2 and RME3, all of which acting on the regulation of meiosis in S. cerevisiae, are similar regarding their function and regulatory mechanisms. IRT1 (IME1 regulatory transcript 1) inhibits the expression of the downstream located and partially overlapping senseoriented gene IME1 (inducer of meiosis 1), thereby preventing erroneous germ cell differentiation and sporulation in haploids (van Werven et al. 2012). In cells grown in a haploid state, the meiosis-repressive transcription factor Rme1 (repressor of IME1) induces the production of the lncRNA IRT1. IRT1 then mediates depositioning of repressive chromatin marks by recruiting the histone methyltransferase Set2 and the histone deacetylase Set3, which leads to a block of transcription initiation of IME1. Moreover, IRT1 hinders binding of the transcription activator Pog1. Upon the onset of meiosis as a response to carbon source and nitrogen starvation in MATa/ MATα heterozygous diploid cells, the expression of the Rme1 encoding gene is inhibited by binding of the diploidspecific a1-α1 repressor complex (Mitchell and Herskowitz 1986). This consequently shuts off formation of the nc transcript IRT1 and allows expression of IME1. Anyway, for the switch from vegetative growth to entry into meiosis in diploids, another mechanism is dominating the regulation of IME1 and IRT1. Under nutrient-rich conditions, PKA (protein kinase A) and TORC1 (target of rapamycin complex I) are produced in both haploid and diploid cells. Those factors are required for binding of the Tup1-Cyc8 complex to the IME1 locus and hence direct inhibition of both IME1 and IRT1 when nutrients are available (Moretto and van Werven 2017). Synergistically, the mechanisms based on regulation by PKA/TORC and Rme1 allow tightly control of entry into meiosis only upon nutrient starvation and in heterozygous diploid S. cerevisiae cells exclusively.  Marina et al. (2013) as antisense, CRF chromatin remodelling factors, IA interaction, intra intragenic, MDED Mmi1-directed exosome degradation, s sense, TGS transcription gene silencing, TF transcription factor Though less research has been done on the two other meiotic lncRNAs in S. cerevisiae, some parallels regarding their regulatory strategies can be noted. Like IRT1, RME2 (regulator of meiosis 2) prevents germ cell differentiation in haploids and interferes with the transcription of its target gene IME4 depending on the presence or absence of Rme1 (Hongay et al. 2006). Also, RME3 (regulator of meiosis 3) represses the expression of its adjacent gene ZIP2, thereby preventing formation of the synaptonemal complex, which is required for chromosome pairing during meiosis (Gelfand et al. 2011). Both RME2 and RME3 are initiated downstream and transcribed in antisense orientation relative to their target genes, and both are interfering with transcription elongation rather than initiation (Gelfand et al. 2011). Furthermore, for both lncRNAs, a mechanism based on regulation of the chromatin status was postulated. However, there are not any details known.
A similar regulatory model as described for IRT1 has also been postulated for the SRG1, a lncRNA that negatively interferes with the expression of the SER3 gene under serine-rich conditions in S. cerevisiae (Martens et al. 2004).
Interestingly, also in S. pombe, meiosis is controlled by a lncRNA termed meiRNA; however, its regulatory strategy is strikingly different. Compared to IRT1 in S. cerevisiae, meiRNA is not involved in the induction of meiosis, but in meiosis progression (Watanabe and Yamamoto 1994) and also chromosome pairing . Two isoforms of the meiRNA (i.e. meiRNA-S and meiRNA-L) are transcribed from the locus and physically interact with RNA-binding protein Mei2 upon onset of meiosis (Watanabe and Yamamoto 1994). During meiosis progression, the meiRNA-Mei2complexes accumulate at the locus, leading to the formation of the so-called Mei2 dot (Yamashita et al. 1998). This finally causes baiting and inhibition of Mmi1, a key-silencing factor associated with heterochromatin formation, RNAi and also Mmi1-mediated RNA degradation (Shichino et al. 2014). As a consequence, genes required for meiosis progression are stably expressed (Fig. 1). Interestingly, due to its ability to physically interact with Mmi1, meiRNA itself is a target of Mmi1-mediated RNA degradation in mitotic cells Yamashita et al. 2012). However, during meiosis, meiRNA is supposed to be stabilized upon complex formation with Mei2 and sequestering of Mmi1 (Harigaya et al. 2006;Hiriart and Verdel 2013). Schematic illustrations of the mechanisms of the here presented lncRNA are provided in several review articles (Hiriart and Verdel 2013;Niederer et al. 2017;Yamashita et al. 2016).
Besides meiosis, a couple of lncRNAs involved in telomere synthesis and maintenance in yeast are known. This process is essential to protect chromosome ends from shortening and prevent erroneous repair initiated by the DNA-damage response (de Lange 2005). In S. cerevisiae, two sorts of telomere-associated lncRNAs have been described: TERRA (telomeric repeat-containing RNA) and the telomerase RNA TCL1. TERRA acts as a scaffold for telomeric DNAs and chromatin-modifying enzymes during telomere synthesis and regulates telomerase activity (Luke et al. 2008). It has been extensively reviewed elsewhere (Cusanelli and Chartrand 2015;Luke and Lingner 2009).
TLC1 provides a platform for the formation for the telomerase complex itself and serves as a template for reverse transcription by Est2 (Zappulla and Cech 2004). A homologue to Two isoforms of meiRNA differing in length result from variation of the polyadenylation sites: meiRNA-L and meiRNA-S. The long version meiRNA-L has the more striking role in meiosis progression. Upon the onset of meiosis, meiRNA-L accumulates with its binding partner Mei2 at the sme2 locus that governs baiting and inhibition of the key-silencing factor Mmi1. Thus, meiosis specific genes, which are destabilized by Mmi1 during mitosis, are stably expressed. Moreover, meiRNA-L mediates chromosome pairing during the meiotic prophase. During mitosis, Mei2 is not produced and meiRNA is destabilized by Mmi1directed exosome degradation. Thus, Mmi1 is active and causes gene silencing by mediating exosome degradation, recruitment of the RNA degradation complexes MTREC and NURS, as well as the RNAi machinery (the histone methyltransferase Clr4 and the RNAi effector complex RITS) and by promoting heterochromatin formation TLC1 was also discovered in S. pombe, namely the telomerase RNA TER1 (Leonardi et al. 2008). Some characteristics are shared between the two telomerase RNAs; however, TER1 was found to be larger than TLC1 and contains a higher number of invariant repeats compared to TLC1 that is rather heterogeneous.
Gene silencing and broad heterochromatin formation is often associated with the degradation of unstable transcripts (reviewed in Tudek et al. 2015;Wu et al. 2012). This issue will be discussed in more detail in BlncRNAs with special properties^.
Also, the response to phosphate starvation in S. pombe is regulated by a set of lncRNAs sharing some features. All of them are transcribed under phosphate-rich conditions and repressed upon starvation and all of them cause repression of their sense-oriented target genes in cis via transcription interference (Ard et al. 2014;Chatterjee et al. 2016;Garg et al. 2018). Nc-tgp1 and prt have been investigated in more detail. In response to extracellular inorganic phosphate, the lncRNA nc-tgp1 is produced and alters the nucleosome density at the promoter of its adjacent gene tgp1 (transporter for glycerophosphodiester 1) (Ard et al. 2014). This results in dissociation of the central transactivator Pho7 and, thus, in a total shut-off of tgp1 transcription (Fig. 2a).
A similar regulatory model was proposed for the regulation of pho1 expression by the lncRNA prt (pho1-repressing transcript). Also in this case, pho1 is repressed in response to prt1 transcription under phosphate-rich conditions and expressed upon phosphate starvation, strictly depending on activation by Pho7 (Fig. 2b) (Chatterjee et al. 2016). Initially, another model for the regulation of pho1 expression by the lncRNA prt was suggested. It is based on the recruitment of Mmi1 by the lncRNA prt, which results in depositioning of repressive chromatin marks and RNAi-mediated silencing (Shah et al. 2014). Yet, later research rather supports the concept of transcription interference by hampering binding of Pho7 (Chatterjee et al. 2016;Garg et al. 2018). Recently, the dissociation of Pho7 from the pho1 locus as a consequence of prt transcription was reported to be governed not only by the lncRNA prt itself but also by RNA polymerase II (Pol II), depending on its phosphorylation status (Chatterjee et al. 2016). According to this model, Pol II moves towards the pho1 promoter during progression of prt transcription and antagonizes binding of Pho7 close to the Poly(A) site of prt, thus resulting in the loss of pho1 initiation (Fig. 2b). Changes in the phosphorylation status of Pol II are supposed to lead to prior termination of prt transcription, consequently resulting in de-repression of pho1. A likewise mechanism was also shown for nc-tgp1 and its target gene tgp1 ). Both lncRNAs, prt and nc-tgp1, are controlled by Mmi1-directed exosome degradation and RNAi, as they harbour a cluster of DSR (determinant of selective removal) motives which are recognized and bound by the central silencing factor Mmi1. Yet, this mechanism is independent from the regulatory impact on their adjacent genes (Ard et al. 2014;Chatterjee et al. 2016).
Recently, a third lncRNA acting on the phosphate metabolism in S. pombe has been discovered, namely prt2 (Garg et al. 2018). Like its two functionally related lncRNAs, prt2 is transcribed upon phosphate starvation and governs repression of its adjacent gene pho84 (Fig. 2b). Moreover, also for prt2, an impact of the phosphorylation status of Pol II on expression of its target gene was confirmed. Conclusively, a similar mechanism as described for prt and nc-tgp can be supposed. Interestingly, prt2 was shown to effect the production of the proximal located lncRNA prt and its regulated gene pho1. Fig. 2 Response to extracellular inorganic phosphate in S. pombe. a Inversely correlated expression of tgp1 and the lncRNA nc-tgp1 in the presence or absence of phosphate. Under phosphate-rich conditions, the lncRNA nc-tgp1 is transcribed and blocks the expression of its sense gene tgp1 by modulation of the local nucleosome arrangement and promoting the dissociation of the central transactivator Pho7. Upon phosphate starvation, nc-tgp1 initiation is prevented, thus allowing binding of Pho7 and expression of tgp1. b Inversely correlated expression of pho84 and the lncRNA prt2 as well as pho1 and the lncRNA prt in the presence or absence of phosphate. Under phosphate-rich conditions, the lncRNA prt2 is transcribed and blocks the expression of its sense gene pho84 by an unknown mechanism. Similarly, the adjacent lncRNA prt is transcribed and blocks the expression of its sense gene pho1 by promoting the dissociation of the central transactivator Pho7. Upon phosphate starvation, prt2 initiation is prevented; thus, Pho84 is produced and in turn acts as a repressor of prt transcription, finally resulting in the expression of pho1 Inactivation of prt2 results in an upregulation of Pho84, which consequently leads to a downregulation of prt and finally a stimulation of pho1 transcription (Garg et al. 2018).

Regulatory strategies of lncRNAs
The simplest and probably most common regulatory strategy of lncRNAs is interference with the transcription of proximal located genes (reviewed in Kornienko et al. 2013;Vance and Ponting 2014). They can govern the expression of sense or antisense located genes in a repressing or activating manner by blocking the transcription machinery, modulation of the nucleosome arrangement and thereby provoking dissociation or binding of regulatory factors like transcription factors. Examples for a repressing effect of sense directed lncRNAs are nc-tgp1, prt and prt2, which have been described in the prior section.
Besides those, an interesting pair of lncRNAs is known that enables cell-cell adhesion during filament formation of S. cerevisiae cells in response to nutrient starvation, namely ICR1 and PWR1 (Bumgarner et al. 2009). They regulate the expression of their adjacent gene FLO11 in a synergetic and sophisticated way. The current concept is a three-state model, comprising an activated, a repressed and a basal state (Bumgarner et al. 2012;Octavio et al. 2009). In the basal state, the lncRNA ICR1 is produced and causes dissociation of the potentially bound activating factor Flo8 as well the repressing factor Sfl1 from the FLO11 promoter. For activation of FLO11 expression, Flo8 is bound and triggers the transcription of the antisense lncRNA PWR1, which acts as a suppressor of ICR1 transcription. In the repressed state, Sfl1 binds to the promoter and mediates silencing by the recruitment of the histone deacetylase Hda1.
Another prominent example for transcription interference by lncRNAs is SRG1, which has been extensively reviewed (Hiriart and Verdel 2013;Niederer et al. 2017;Yamashita et al. 2016) and was briefly mentioned in the prior paragraph. It is transcribed under serine-rich conditions in response to activation by Cha4 (Martens et al. 2005) and mediates nucleosome depositioning at the SER3 promoter depending on FACT, Spt6/Spn1 and Spt2 in order to repress the expression of SER3 (Martens et al. 2004). As depicted in the prior section, similar mechanisms were described for IRT1, RME2 and RME3, yet in this case, also an impact on the chromatin arrangement by directing depositioning of repressive chromatin marks was postulated.
The quite newly identified pHO-lncRNA forces nucleosome repositioning at the locus of the downstream located HO gene in S. cerevisiae (Yu et al. 2016). The regulated gene encodes the HO endonuclease, which is responsible for mating type interconversion during re-entering of the cell cycle after pheromone-dependent arrest in G1. Transcription of pHO-lncRNA is induced in response to the production of a pheromone (i.e. the α-factor) and causes nucleosome rearrangement and displacement of the factor SFB from the HO promoter. The loss of the activating signal from SFB results in a shut-off of HO expression and thus prevention of matingtype interconversion during re-entering into the cell cycle.
An exceptional mechanism for the regulation by a lncRNA was recently described for SUT169 (Huber et al. 2016). Upon nutrient starvation, SUT169 is transcribed and promotes the production of a sporulation-specific protein encoded by the SPS100 gene. However, in contrast to other lncRNAs, SUT169 does not activate the expression of it target gene, but it interferes with the transcription by supporting the production of a longer and more stable isoform of SPS100.
Another regulatory strategy followed by lncRNAs is depositioning of histone modifications and the recruitment of chromatin remodelling factors. One well-known example for this is the GAL10 lncRNA. In S. cerevisiae, the expression of the GAL genes (i.e. GAL1 and GAL10) is switched on in the presence of galactose and repressed in the presence of glucose to allow regulation of galactose metabolism. Under repressing conditions, GAL10 ncRNA is produced and mediates di-and trimethylation of K4 and dimethylation of K36 on histone 3 by Set2 (Houseley et al. 2008). Those repressive chromatin marks are bound by Eaf3, which recruits the histone deacetylase Rpd3S, thus resulting in broad deacetylation and silencing of the whole GAL locus. In addition, H3K4me2 and H3K4me3 cause a delay in the recruitment of Pol II and TBP to the GAL10 promoter (Geisler et al. 2012). Another lncRNA acting on the GAL genes is GAL4 lncRNA. Here, more studies are required to understand its mechanism in detail (Geisler et al. 2012).
Further examples of lncRNAs effecting the chromatin organization are the mlonRNA from S. pombe and the ncASP3 and the antisense lncRNA of the CDC28 gene in S. cerevisiae, all of which are involved in stress response. The latter is induced by the stress-activated protein kinase Hog1 upon osmostress and supports translocation of Hog1 to the overlapping gene CDC28 by bending the local genomic region into a loop (Nadal-Ribelles et al. 2014). This results in the recruitment of chromatin remodelers and thus activates the expression of the cyclin-dependent kinase 1 encoding gene CDC28.
Similarly, glucose starvation in S. pombe triggers the production of noncoding transcripts upstream of the fbp1 gene, the so-called mlonRNAs (Hirota et al. 2008). These metabolic stress-induced lncRNAs trigger a switch of the chromatin configuration to an open state, which makes DNA more accessible for binding of Pol II and activators such as Atf1 (Hirota et al. 2008;Takemata et al. 2016). As a consequence, the fbp1 gene (encoding the fructose-1,6-bisphosphatase) is expressed in the absence of glucose. The production of mlonRNAs in S. pombe is anticorrelated with antisense transcripts, which are produced from the fbp1 locus under glucose-rich conditions (Miki et al. 2016). Compared to the fbp1 mRNA, both mlonRNAs and their antisense transcripts were found to be prone to degradation by the nuclear exosome/Rrp6 complex (Galipon et al. 2013;Miki et al. 2016).
The lncRNA ncASP3 in S. cerevisiae influences the expression of its target gene ASP3 in a peculiar way. ASP3 is regulated by nitrogen catabolite repression (NCR) and encodes asparaginase II, an enzyme performing hydrolysis of asparagine to aspartate and ammonium cations in response to nitrogen starvation (Dunlop et al. 1978). ncASP3 is expressed in both cases, when nitrogen is available and upon nitrogen depletion, and maintains an open chromatin configuration at ASP3 by mediating trimethylation of H3K4. Thus, ASP3 is kept in an accessible status, allowing immediate expression upon the onset of NCR (Huang et al. 2010).
Another type of regulatory strategy of lncRNAs is the interaction with proteins. lncRNAs can attract proteins, affect their function or act as scaffolds for several factors. Welldescribed examples for this in yeast are meiRNAs, TERRA or telomerase RNAs. They have been presented in detail in the prior section. Also for the recently identified lncRNA HAX1 in T. reesei, an interaction with the main transactivator of cellulase expression can be supposed as a sequence element rich in Xyr1-binding sites is present at its 5′ end (Till et al. 2018). To date, details regarding the regulatory strategy of HAX1 have not been reported.
lncRNAs with special properties Some lncRNAs are unstable transcripts (reviewed in Garneau et al. 2007;Tudek et al. 2015). They can be destabilized by different cellular components; anyway, their degradation allows tight control of their regulatory impact. One type of unstable lncRNAs are CUTs, which are degraded in the nucleus by the exosome/Rrp6 complex (Davis and Ares Jr 2006;Wyers et al. 2005). Examples are the PHO84 antisense transcript (Camblong et al. 2007) or TY1 (Berretta et al. 2008). As described for prt and nc-tgp1, degradation by the nuclear exosome can also be triggered by Mmi1 (Harigaya et al. 2006). XUTs and SUTs in contrast are exported to the cytoplasm for processing. There, they might be destabilized by decapping mediated by Dcp1 and Dcp2 (Garneau et al. 2007). XUTs then are degraded by the 5′-3′ exonuclease Xrn1 (Nagarajan et al. 2013) or the S. pombe ortholog Exo2 (Szankasi and Smith 1996), whereas SUTs are processed by the cytoplasmic exosome (Garneau et al. 2007). Several XUTs and SUTs were shown to be targeted by the nonsensemediated decay pathway, which is a cytoplasmic surveillance mechanism for recognizing and discarding mRNAs with premature stop codons Tudek et al. 2015;Wery et al. 2016). One example for XUTs is the GAL10 lncRNA (Houseley et al. 2008). NUTs (Nrd1-undetermined transcripts) are not controlled by post-transcriptional degradation but are sensitive to prior termination of lncRNA transcription by Nrd1 (Schulz et al. 2013) or the S. pombe ortholog Seb1 (Mitsuzawa et al. 2003). All these types of unstable transcripts cause silencing of their target genes. The mechanisms of degradation described here arise as a control of pervasive transcription and often act as a protection from gene silencing throughout the life cycle (Wyers et al. 2005). Hence, one could question whether the production of these unstable transcripts is really an explicit mechanism or rather an unwanted side effect of exhaustion of transcriptome surveillance.
Most lncRNA have a repressing effect on their target genes. The few examples of fungal lncRNAs acting as activators comprise the CDC28 antisense transcript, meiRNA, SUT169, SUT457, SPNCRNA.1164 and HAX1. Yet, in fact, they hardly share any features: some act in cis, others in trans; they occur in different organisms and are involved in different cellular processes ranging from telomere control over sporulation to metabolism and stress response. Details regarding the regulatory mechanisms of SUT457, SPNCRNA.1164 and HAX1 are unknown. However, due to its outstanding role as the first characterized lncRNA in a filamentous fungus, we will further address to HAX1 in this chapter.
HAX1 was incidentally discovered by random integration of a marker cassette into the genome of T. reesei (Till et al. 2018). T. reesei is a saprophyte, growing on dead plant material (Klein and Eveleigh 1998). It secretes large quantities of cellulases and hemicellulases, which cause degradation of complex plant biopolymers (Ouyang et al. 2006). Especially, cellulases are commonly used for several processes meeting human regards; hence, T. reesei is widely applied as a highyield producer of those enzymes in industry (Kubicek and Penttilä 1998). The lncRNA HAX1 was identified as an activator of cellulase expression (Till et al. 2018). Thus, HAX1 has a pivotal role, both as a player in the complex regulatory network of cellulase expression and as a potential biotechnological tool for the improvement of enzyme production. While details on the regulatory strategy are not reported yet, the presence of a sequence element rich in Xyr1-binding sites suggests an interplay with the main transactivator of cellulase and xylanase expression, the Xylanase regulator 1 (Xyr1) (Rauscher et al. 2006;Stricker et al. 2006). One special feature regarding hax1 expression is the strain-specific variation of RNA length (Till et al. 2018). Interestingly, much shorter versions of HAX1 are present in the T. reesei wild-type strain than in moderate or in high cellulase overproduction strains. Overexpression of the three identified HAX1 versions in the wild-type strain led to an increase in cellulase activity, depending on RNA length. The shortest version only slightly improved cellulase expression, whereas the longest version led to the strongest increase in cellulase activity (Fig. 3). Conclusively, a direct link between HAX1 length and its regulatory role can be supposed.
Another example for the variation of transcript length in fungi is the above-discussed meiRNA in S. pombe. Two isoforms of this lncRNA are produced: the 0.5 kb meiRNA-S and the approximately 1.2 kb meiRNA-L (Watanabe and Yamamoto 1994). The production of these isoforms depends on the presence of different polyadenylation sites, initiating transcription termination (Fig. 1). In contrast, the differences in HAX1-length result from variation of the transcription start point (Fig. 3). Also, the transcription of meiRNA-L is initiated at a more or less defined region rather than a certain position. However, in literature, the variety of transcripts of meiRNA-L slightly differing in length is specified as one RNA species. A vague definition of transcription start-and end points is not uncommon for lncRNAs. Heterogeneity in 3′ and 5′ ends was also stated for other lncRNAs such as telomerase RNAs (Dandjinou et al. 2004;Leonardi et al. 2008). However, the special feature of meiRNA-and HAX1 isoforms is that they have different regulatory impacts. Similar to HAX1 and against initial assumptions, the long isoform meiRNA-L turned out to have a more striking role in meiosis progression than meiRNA-S, both for the attraction of Mmi1 (Shichino et al. 2014) and for chromosome pairing . Anyway, strain-specific variation of RNA length remains an exceptional strategy solely described for HAX1 so far.
As the functionality of lncRNAs commonly depends on their distinct folding, grouping of lncRNAs based on their structural features might also be a promising way to go. This would also allow the separation of certain species such as unstable transcripts (e.g. CUTs, XUTs, NUTs) or ribonucleoproteins (e.g. telomerase RNAs). In human, attempts for the prediction of lncRNAs based on structural mappings were presented (Washietl et al. 2005a;Washietl et al. 2005b). Also in yeast, studies on genome-wide profiling of the secondary structures of ncRNAs were performed (Kertesz et al. 2010). Yet, more research would be needed to better assess the potential of this feature for the classification of lncRNAs.

Summary and conclusion
lncRNAs emerged as a heterogeneous group of noncoding transcripts involved in a variety of regulatory processes in eukaryotes. Initially, they were discovered in human; meanwhile, several lncRNAs have been identified in fungi. Some examples of yeast lncRNAs are well known and thoroughly described in literature, whereas others have just recently been identified and need to be investigated in more detail. lncRNAs are very diverse regarding their features, functions and regulatory strategies. They are involved in cellular processes such as metabolism and stress response, cell cycle control (e.g. meiosis), telomere maintenance or gene silencing. Their regulatory mechanisms range from transcription interference and local nucleosome rearrangement, over depositioning of histone modifications and chromatin remodelling, to physical interactions with several factors, acting as a scaffold or bait. Some lncRNAs might be distinguished upon certain characteristics (e.g. CUTs, XUTs, NUTs), and others acting on the same cellular process share a similar mode of action. However, in general, classification of lncRNAs is a challenging task. One possible property for categorizing lncRNAs in the future might be their higher ordered structure. Unfortunately, at the current state of knowledge, not enough data are available on this issue.
In addition to yeast lncRNAs, recently, the first functional characterization of a lncRNA was reported for a filamentous fungus, namely T. reesei. It was presented as a potential tool for strain improvement and the industrial exploitation of T. reesei. Similarly, a specific use of yeast lncRNAs for industrial purposes might be considered. Both S. cerevisiae and S. pombe are commonly applied for biotechnological production processes. Hence, targeted intervention in processes Fig. 3 Regulation of cellulase gene expression by the lncRNA HAX1 in T. reesei. Three isoforms of HAX1 differing in length result from variation of the transcription start point in different T. reesei strains: HAX1 QM6a , HAX1 QM9414 and HAX1 Rut-C30 . They act as activators of cellulase expression. The longest version HAX1 Rut-C30 has a higher impact on the cellulase activity compared to HAX1 QM9414 and HAX1 QM6a (indicated by triple, double and single plus symbols, respectively). The regulatory mechanism of HAX1 is unknown, yet an interplay with the main transactivator Xyr1 is supposed. For details, see text influenced by lncRNAs represents a promising strategy for process optimisation. Ongoing research and steadily improving technologies provide the basis for uncovering yet unknown mechanisms of lncRNAs, their biotechnological applications and the identification of new candidates in yeasts and higher fungi.
Funding Information Open access funding provided by Austrian Science Fund (FWF). This work was supported by a grant from the Austrian Science Fund (FWF): [P26618]. Further financial support by Novozymes, by the Austrian Federal Ministry of Science, Research and Economy and by the National Foundation for Research, Technology and Development is gratefully acknowledged.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.