La, NF1 family proteins, PC4 and topoisomerase I have been proposed as accessory pol III transcription factors that facilitate multi-cycle transcription by hspol III. hsLa is an RNA-binding protein, with an N-terminal segment that binds to the short 3' terminal U tracts of primary pol III transcripts and a C-proximal site that recognizes primary transcript 5' ends. Phosphorylation of C-proximal Ser366 is required for La's participation in tRNA maturation (Maraia and Intine, 2001). In a Human Polymerase III in vitro system for transcriptional recycling that uses arrested transcription elongation complexes on immobilized DNA templates, further rounds of transcription require La (Maraia, 1996) and the just referred-to C-proximal Ser phosphorylation is required for this function (Fan et al., 1997). In contrast, highly efficient (>99%) immunodepletion of La from a frog oocyte extract that executes very efficiently recycling transcription in vitro has no effect on transcript yield (LinMarq and Clarkson, 1998). Although La is detected as a component of human holoPolymerase III, the gene encoding the yeast homologue of La is not essential (Yoo and Wolin, 1994). Thus, La is unlikely to be an essential and general transcription-recycling factor of pol III. Whether it is a significant factor in human pol III transcription remains unresolved (Maraia and Intine, 2001).

Three other components that co-purify hsholoTFIIIC (Schramm and Hernandez, 2002) have been implicated in transcriptional termination and recycling: PC4, topoisomerase I and NF1. Topoisomerase I and PC4 were shown to promote multi-cycle transcription selectively and separately (Wang and Roeder, 1998), but the physiological significance of their activity has been discounted subsequently on account of the very great molecular excess of protein required for this action (Wang et al., 2000). Purification of a DNA-binding component of a TFIIIC-containing fraction (Wang and Roeder, 1998) to apparent homogeneity yields NF1 family proteins that bind to the adenovirus VAI gene's terminator. A recombinant NF1 protein also interacts with two subunits of hs TFIIIC2. The biochemically purified NF1 family proteins exert strong effects on transcriptional termination and increase the number of rounds of transcription in vitro. Detailed analysis of this effect has been confined to the VAI gene, which harbors consensus NF1 binding sites that overlap its two terminators and are essential for the transcription termination activity of NF1. However, terminator-overlapping NF1 sites are not a general feature of hspol III-transcribed genes. Involvement of NF1 proteins in transcription of genes lacking NF1-binding sites, conceivably through interaction with TFIIIC2, has been referred to (Wang et al., 2000) but not presented in concrete detail. [Nor has the implicit conflict with the proposed essential character of NF1-binding sites at the VAI gene terminator been resolved.] Biochemically purified NF1 has been reported to exert no effect on transcription in a highly purified system consisting only of recombinant TFIIIB, immunopurified TFIIIC and pol III (Wang et al., 2000).

Because pol III transcription units are small, the transcribing polymerase never leaves the general vicinity of its promoter; DNA bending by transcription factors further compacts these transcription units. As a consequence, repeated utilization of an (isolated) pol III gene by the same molecule of polymerase could be based on purely stochastic principles. In addition, evidence for the operation of a specific handing-off mechanism that repetitively returns pol III from its termination site to the (same gene's) promoter has been presented (Dieci and Sentenac, 1996). However, the task of discriminating between purely stochastic recycling and handing-off mechanisms involving specific protein-protein interactions presents difficult technical challenges that have not yet been met. (Some of these difficulties are briefly discussed in (Geiduschek and Kassavetis, 2001).) Thus, the existence of a mechanism for directly guiding transcription-terminating pol III back to the same promoter (i.e. in cis), and the nature of the protein-protein interactions that might generate such a chaperone-like mediation remain to be established.