Transgenes and protein localization: myths and legends Francesca M. Quattrocchio, Cornelis Spelt, and Ronald Koes Department of Molecular Cell Biology, Graduate School of Experimental Plant Sciences, VU-University, Amsterdam, The Netherlands Fluorescent protein (FP) fusions are frequently used to localize and follow the movement of proteins in living cells. However, a consensus is missing about the experi- mental design and controls that guarantee the reliability of the results. Here, we discuss possible artifacts and try to navigate through the many methods, preferences, and assumptions that surround protein localization in plants that make it difficult to design a universal ap- proach to achieve reliable results. Protein localization studies We moved recently into plant cell biology and, as new- comers, we were amazed at the wide range of standards regarding the use of FP fusions. It is sometimes taken for granted that FP fusions faithfully report the trafficking and localization of native proteins, whereas in other cases dire warnings are given reminding one that FP fusions might accumulate differently compared with the native protein, because of overexpression or misfolding [1]. We used FP fusions of numerous proteins and obtained convincing and reproducible localization results as well as artifacts that provide valuable lessons. Here, we discuss several technical aspects of protein localization studies, and describe some artifacts that often go unrecognized with the most commonly used controls and that may have led to flawed studies. Expression driven by the CaMV35S promoter The 35SRNA promoter (35S) of cauliflower mosaic virus is widely used to express transgenes, including FP fusions. To avoid artifacts generated by overexpression, some researchers advocate using the promoter from the endo- gene. We agree that overexpression may cause mislocali- zation of FP fusions, but contest that the use of the 35S promoter is a common source of such artifacts. The 35S promoter comprises distinct enhancers that collectively result in (nearly) constitutive expression [2]. Hence, RNAs expressed from 35S are abundant in RNA samples from entire organs, compared with RNAs from endogenous genes, due to expression in (almost) all cells and not because expression in individual cells is high. By contrast, mRNAs from endogenous genes are usually re- stricted to a limited number of cells, which makes it diffi- cult to assess their abundance per cell from the amount of mRNA in an entire organ. Flower pigmentation genes provide a nice example, because petals comprise only a colored epidermis, expressing pigmentation genes, and a few layers of uncolored mesophyll cells [3,4]. 35S-driven transgenes rescue pigmentation mutants only in those transformants with the highest transgene expression (e.g. [4–6]), which is usually not much higher than expression of the endogene (Figure 1A). Because the endogene is primari- ly expressed in the adaxial epidermis (Figure 1B) and 35S- driven transgenes in most cells [2], the abundance of trans- gene-derived RNAs within epidermal cells is at best similar, but often lower, than endogene-derived RNAs. The flower meristem identity gene DOUBLE TOP (DOT) provides an extreme example. DOT mRNA is highly expressed, giving strong in situ hybridization signals, in a few cells between sepal and petal primordia; therefore, it is difficult to detect by RT-PCR in inflorescence apices [7]. By contrast, 35S:DOT transcripts are easily detected by RT- PCR, but not by in situ hybridization, and also fail to restore floral organ identity in dot mutants [7]. Experi- ments with the syntaxin gene KNOLLE (KN) and 35S:KN transgenes in Arabidopsis (Arabidopsis thaliana) gave similar results [8]. These examples show that a 35S-driven transgene is not ‘overexpressed’ (at a higher level than the endogene within a certain cell), but is expressed ‘ectopical- ly’ (outside the normal domain) or ‘constitutively’ (contin- uously in all cells). Other factors, such as translation efficiency and protein stability, are ignored in this discussion. For example, frameshift mutations adding different tails to the tran- scription factor ANTHOCYANIN1, result in protein ex- pression levels differing >100-fold [9]. Hence, fusion of an FP tag may affect protein accumulation more dramatically than might the choice between 35S or native promoters, yet this is rarely considered. When is an FP fusion a good marker? Journals often demand the reliability of FP fusions be demonstrated by mutant complementation [1]. However, this does not guarantee that both the protein and FP tag arrived at the correct destination, because they can be separated by cleavage and have different fates. Analysis of various GFP fusions indicates this to be a frequent cause of unreliable results. The MYB transcription factor ANTHOCYANIN2 (AN2) of Petunia (Petunia hybrida), for example, activates antho- cyanin synthesis in petals and AN2 expression from the 35S promoter rescues pigmentation of an2 petals and also pig- ments leaves and stems [5]. Constitutive expression of AN2 fusions with an N- or C-terminal GFP tag (35S:GFP-AN2 Techniques & Applications Corresponding author: Quattrocchio, F.M. (f.m.quattrocchio@vu.nl). Keywords: transgenes; CaMV35S promoter; GFP fusions; protein localization. TRPLSC-1078; No. of Pages 4 1