The results of this study demonstrate that two NAT1 variants, *11 and *10, account for most of the observed genetic influence on NAT1 protein expression. Both *10 and *11 represent gain of function polymorphisms, leading to increased protein level or enzyme activity in livers and B-lymphocytes, thereby resolving previous contradictory results [23
] . While *11 was shown to direct preferential polyA2 site usage, yielding more of the major 3′UTR isoforms endowed with higher translation efficiency, the regulation of *10 involves increased protein translation efficiency without affecting mRNA levels or processing. Moreover, the association between NAT1 *10 and *11 and SMX-induced hypersensitivity in HIV/AIDS patients further demonstrates the gain-of-function of NAT1 *10 and *11. With these results, we classify *4/*4 carriers as reference acetylators, *4/*10 carriers as intermediate acetylators, and *10/*10 and *11 carriers as fast acetylators, reflecting the stronger increase of protein level or enzyme activity caused by *11 compared to *10 (). Homozygous *11/*11 carriers could be considered ultra-fast metabolizers but are rare (
NAT1 *11 and *10 have been implicated as risk alleles in several clinical association studies [12
], mostly being referred to as fast acetylator alleles [15
], while the biological mechanisms remained uncertain. Our finding that *11 increases NAT1 protein level/enzyme activity involving switching of polyA site usage was supported by both in vivo
and in vitro
results: (1) Allelic mRNA analysis, in vitro
NAT1 expression, and 3′UTR reporter gene assays demonstrate that *11 yields higher levels of the major 3′UTR NAT1 isoform . (2) The major 3′UTR isoform displayed higher protein translation efficiency than the short isoform. (3) NAT1 immunoreactivity and enzyme activity is higher in *11 than *4 liver or B-lymphocytes. Consistent with this, *11 had been associated with higher enzyme activity in previous association studies [12
] and in this study.
mRNA polyadenylation is a cellular process that adds polyA tails to the vast majority of maturing mRNAs, affecting mRNA stability, translation, and transport [52
]. mRNA polyadenylation starts with an endonucleolytic cleavage at a site determined by the surrounding RNA sequence (cis
-elements) and their binding proteins (trans
factors). PolyA addition is determined by a polyadenylation signal, AAUAAA or a close variant located 10–35 bp upstream of the cleavage site, acting in concert with auxiliary upstream and downstream elements [48
]. The switch from polyA site 1 to polyA site 2 usage in *11, observed in both liver and B-lymphocytes, appears to be determined by the 9bp deletion located near the polyA1 site, as well as DNA sequence downstream of the polyA2 site, including three SNPs (rs28359534, rs8190862 and rs8190863) in complete LD with *11, but not previously included with this haplotype. This potential long-range conformational folding effect was suggested by 3′UTR reporter gene results showing that only the reporter gene constructs carrying the long 3′UTR isoform resulted in greater luciferase activity for *11 than *4, whereas reporter genes with short and major 3′UTRs showed no difference between *11 and *4 (). Moreover, NAT1 expression of cDNA lacking sequence downstream of polyA2 site failed to show difference in NAT1 mRNA expression and enzyme activity between *4 and *11. All three NAT1 3′UTR isoforms appear to have similar turn-over rates, because allelic expression imbalance was undetectable after PCR amplification of total NAT1 mRNA encompassing all isoforms. Therefore, *11 selectively increased usage of the polyA2 site resulting in increased levels of the major 3′UTR isoform, and hence increased protein translation. Taken together, the *11 allele causes a robust gain-of-function.
Coding region SNPs can directly affect protein sequence and function (nonsynonymous), or mRNA processing and translation (synonymous and non-synonymous), as we had reported previously [45
]. The results in the present study indicate that three coding region SNPs present in *11 did not change mRNA expression nor protein level/protein activity. This result is consistent with two previous reports [13
], but inconsistent with others [24
]. Because regulation of *11 involves switching of polyA site usage, changes in NAT1 activity of *11 cannot be observed after transfection of coding region sequence lacking the full length 3′UTR. However, we cannot rule out the possibility that the non-synonymous coding region SNPs in *11 change NAT1 activity for some substrates [24
], other than PABA used in the present study. Also, we cannot exclude the possibility that the non-synonymous SNP embedded in *11 affects enzymatic activity with respect to SMX, as the enzyme activity was measured here with PABA as a substrate. Thus far, there is no clear evidence in the literature that *11 affects substrate selectivity, but more detailed studies should be carried out to address this possibility.
Our results also differ from a previous report showing lower or unchanged of NAT1 enzyme activity of *11 and *10 in whole blood lysates [22
]. While regulatory polymorphisms could be tissue specific as we have demonstrated for VKORC1 and CYP3A4 [26
], our results suggest the regulation of NAT1 *10 and *11 are similar in livers and B-lymphocytes. However, we cannot rule out different regulation of other cells in whole blood. Moreover, since NAT1 displayed a wide range of inter-person variability in total mRNA expression (>10 fold) in B lymphocytes and livers independent of cis
-acting polymorphisms, other factors, for example, epigenetic regulation or trans-acting factors may mask or even override the effects from cis
-acting polymorphisms. It is further important to note that the most abundant heterozygous carriers of *10 are not expected to show a pronounced change in NAT1 activity, as demonstrated here.
Two possible mechanisms had been advanced previously underlying NAT1 *10 regulation: (1) altered consensus polyadenylation signal (AAT
AAA to AAA
AAA) leading to changes in polyadenylation, and enhanced mRNA stability [7
]; (2) interaction of an RNA-binding protein with NAT1 mRNA, thereby increasing NAT1 mRNA stability[55
]. Both of these two possible mechanisms would have resulted in increased NAT1 mRNA levels associated with *10, and therefore, in detectable allelic expression imbalance. However, the present results demonstrate that *10 did not yield changes in mRNA levels (Supplemental Figure 4
), nor did it change the usage of polyA sites (). In contrast, 3′UTR reporter gene assays indicate that NAT1 *10 increases protein translation efficiency (). The mechanism underlying this regulation remains unclear. Since *10 resides close to a microRNA target site, it is possible that microRNAs may be involved in *10 regulation, requiring further investigation.
We also have observed differences in tissue-selective expression of 5′UTR splice variants, and inter-individual variability in type II 5′UTR splice variants in B-lymphocytes. Since 5′UTR variants can affect protein translation efficiency [3
] , the variation in 5′UTR splicing may impose another source of tissue-selective inter-individual variability in NAT1 enzyme activity. However, our result suggests that inter-person variability of type II transcripts was unrelated to genetic factors, and clearly independent of *10 and *11.
While the main focus of this study was on molecular genetics of NAT1 *10 and *11, we also explored the in vivo
impact of *10 and *11 on SMX metabolism/inactivation. N-acetylation mediated by NAT1 and NAT2 are the major inactivation pathway for SMX [32
]. Because NAT1 has higher binding affinity and Vmax for SMX [33
] and has broader tissue expression than NAT2, NAT1 may be the primary determinant of SMX clearance both systemically and in target tissues, such as skin and immune-responsive cells. However, the association between NAT1 polymorphisms and SMX-induced hypersensitivity had not been extensively studied, because the function of NAT1 *10 and *11 was previously uncertain, and NAT1 is often considered to be monomorphic. The molecular genetics results reported here led to the prediction that NAT1 *10 and *11 may be protective against SMX-induced hypersensitivity. In contrast to loss of function polymorphisms in NAT2, the effect of which appears to be confounded by reduced liver enzyme activity caused by HIV infection [40
], gain of function polymorphisms could have a persistent effect even in HIV/AIDS subjects. Our result indicate for the first time that NAT1 fast acetylators (*10/*10 and *11 carriers) are protected against SMX-induced hypersensitivity in HIV/AIDS, but only in slow NAT2 metabolizer subjects. While this example is one of a few studies revealing a robust gene-gene interaction, our results require replication in an independent patient cohort.
In summary, we have identified *11 and *10 as functional alleles that increase NAT1 enzyme activity through increasing protein translation efficiency. Since NAT1 *11 and *10 have been associated with various cancers and with response to drugs and xenobiotics, understanding the molecular mechanisms will provide a solid foundation for clinical association studies and assist in the design of biomarkers for personalized medicine.